From dimitri.lecas at c-s.fr Tue Oct 2 04:14:54 2007 From: dimitri.lecas at c-s.fr (LECAS Dimitri) Date: Tue, 02 Oct 2007 11:14:54 +0200 Subject: Solve a symmetric matrix Message-ID: <13d0fb13f394.13f39413d0fb@c-s.fr> Hi, I try to solve a symmetric problem, i use MatCreateMPISBAIJ for the matrix but i have an error message about the preconditioning (PCILU). With the option -pc_type none, i have big problem with convergence (residual > 1e32). So my question is what the option (for ksp and pc) to use for symmetric problem. I not sure i understand the meaning of block size in MatCreateMPISBAIJ, i set it to 1. Best Regards. -- Dimitri Lecas From hzhang at mcs.anl.gov Tue Oct 2 08:51:47 2007 From: hzhang at mcs.anl.gov (Hong Zhang) Date: Tue, 2 Oct 2007 08:51:47 -0500 (CDT) Subject: Solve a symmetric matrix In-Reply-To: <13d0fb13f394.13f39413d0fb@c-s.fr> References: <13d0fb13f394.13f39413d0fb@c-s.fr> Message-ID: PCILU and PCICC are not supported for MPISBAIJ matrix type (the error message should indicate this problem). For sequential sbaij matrix, you can use PCICC. For parallel, you may try the default petsc preconditioner block jacobi + icc. See Chapter 4 of petsc user manual. You can start testing a petsc example ~petsc/src/ksp/ksp/examples/tutorials/ex5.c with runtime options: mpirun -np 2 ex5 -mat_type sbaij -mat_ignore_lower_triangular -ksp_view On Tue, 2 Oct 2007, LECAS Dimitri wrote: > Hi, > > I try to solve a symmetric problem, i use MatCreateMPISBAIJ for the > matrix but i have an error message about the preconditioning (PCILU). > With the option -pc_type none, i have big problem with convergence > (residual > 1e32). So my question is what the option (for ksp and pc) to > use for symmetric problem. > > I not sure i understand the meaning of block size in MatCreateMPISBAIJ, > i set it to 1. When set the block size=1, you acturally use a symmetric matrix that only stores upper triangular part. Hong > > Best Regards. > -- > Dimitri Lecas > > From bsmith at mcs.anl.gov Tue Oct 2 19:53:29 2007 From: bsmith at mcs.anl.gov (Barry Smith) Date: Tue, 2 Oct 2007 19:53:29 -0500 (CDT) Subject: Preallocation for external direct solvers In-Reply-To: <46213F12.9050906@ices.utexas.edu> References: <46213F12.9050906@ices.utexas.edu> Message-ID: Paul, Unfortunately there is no good fix for using the MatCreateSeqAIJ() directly. You will need to use MatCreate() MatSetFromOptions() or MatSetType(mat,MATSUPERLU_DIST,ierr) MatSeqAIJSetPreallocation(....) MatMPIAIJSetPreallocation(....) Note I call both so it will work properly on 1 or more processes. You MUST call the MatSeqAIJSetPreallocation() AFTER the matrix type has been set. I have tried to update the manual pages for the matrix types to clarify this for future users. Thank you for pointing out the problem. Barry On Sat, 14 Apr 2007, Paul T. Bauman wrote: > Hello, > > I'm developing some code and am using direct solvers (for the moment), in > particular MUMPS and SuperLU. I noticed that PETSc is not preallocating the > memory for it even though I specify maximum number of entries per row. > > call MatCreateSeqAIJ(PETSC_COMM_WORLD,total_num_dofs,total_num_dofs, & > max_number_nonzeros_row,PETSC_NULL_INTEGER, J ,ierr) > > call MatSetFromOptions(J,ierr) > > call SNESSetJacobian(snes,J,J,FormJacobian,PETSC_NULL_OBJECT,ierr) > > When I print -info: > > [0] User provided function(): (Fortran):PETSc successfully started: procs 1 > [0] User provided function(): Running on machine: dhcp-67-30.ices.utexas.edu > [0] PetscCommDuplicate(): Duplicating a communicator 1140850688 -2080374784 > max tags = 2147483647 > [0] PetscCommDuplicate(): Using internal PETSc communicator 1140850688 > -2080374784 > [0] PetscCommDuplicate(): Using internal PETSc communicator 1140850688 > -2080374784 > [0] PetscCommDuplicate(): Using internal PETSc communicator 1140850688 > -2080374784 > [0] PetscCommDuplicate(): Using internal PETSc communicator 1140850688 > -2080374784 > [0] MatConvert_AIJ_AIJMUMPS(): Using MUMPS for LU factorization and solves. > [0] PetscCommDuplicate(): Using internal PETSc communicator 1140850688 > -2080374784 > 0 SNES Function norm 8.500000000000e-01 > [0] MatSetUpPreallocation(): Warning not preallocating matrix storage > > So, of course, this really slows down the code for even small problems (20000 > dof). Same story for SuperLU. Should I use a different call to do the > preallocation? > > Thanks for any suggestions. > > Best, > > Paul > > From joaoflavio.vasconcellos at jcu.edu.au Wed Oct 3 00:11:59 2007 From: joaoflavio.vasconcellos at jcu.edu.au (JoaoFlavio Vasconcellos) Date: Wed, 3 Oct 2007 15:11:59 +1000 (EST) Subject: SNES problem Message-ID: <20071003151159.BMZ12167@mirapoint-ms1.jcu.edu.au> Hello I'm trying to use SNES to solve a problem. I wrote the code above to evaluate the residual of the linear system. In this routine norm_2 of F is different from zero but outside of this routine norm_2 of F is zero. #undef __FUNCT__ #define __FUNCT__ "FormFunction1" PetscErrorCode FormFunction1(SNES snes, Vec X, Vec F, void *_simulator) { Simulator *simulator = (Simulator*) _simulator; PetscErrorCode ierr; VecView (X, 0); ierr = simulator->CalculateMatrixCoefficient(X); CHKERRQ(ierr); ierr = simulator->CalculateSourceTerm(X); CHKERRQ(ierr); ierr = simulator->CalculateResidual (X, &F); CHKERRQ(ierr); cout << "Residual " << endl; VecView (F, 0); // norm_2 of F is != zero return 0; } // End of routine Messages from -snes_view: 0 SNES Function norm 0.000000000000e+00 SNES Object: type: ls line search variant: SNESLineSearchCubic alpha=0.0001, maxstep=1e+08, steptol=1e-12 maximum iterations=50, maximum function evaluations=10000 tolerances: relative=1e-08, absolute=1e-50, solution=1e-08 total number of linear solver iterations=0 total number of function evaluations=1 KSP Object: type: gmres GMRES: restart=30, using Classical (unmodified) Gram-Schmidt Orthogonalization with no iterative refinement GMRES: happy breakdown tolerance 1e-30 maximum iterations=10000, initial guess is zero tolerances: relative=1e-05, absolute=1e-50, divergence=10000 left preconditioning PC Object: type: ilu ILU: 0 levels of fill ILU: factor fill ratio allocated 1 ILU: tolerance for zero pivot 1e-12 out-of-place factorization matrix ordering: natural // end of message I dont know why "SNES Function norm 0.000000000000e+00" if F's norm_2 inside my routine is not. Has anyone that could sugest how to fix this problem ? Thank you in advance. Joao Flavio Vieira de Vasconcellos School of Engineering James Cook University Phone: +61 7 4781 4340 Fax: +61 7 4775 1184 Email: jflavio at iprj.uerj.br joaoflavio.vasconcellos at jcu.edu.au skype: jflavio_vasconcellos msn: jf_vasconcellos at hotmail.com From dalcinl at gmail.com Wed Oct 3 10:40:28 2007 From: dalcinl at gmail.com (Lisandro Dalcin) Date: Wed, 3 Oct 2007 12:40:28 -0300 Subject: SNES problem In-Reply-To: <20071003151159.BMZ12167@mirapoint-ms1.jcu.edu.au> References: <20071003151159.BMZ12167@mirapoint-ms1.jcu.edu.au> Message-ID: Flavio, are you completelly sure your FormFunction1() is actually being called? I can see the SNES monitor at iter 0, but I did not see in the the outputs of VecView() or your 'cout << "Residual " << endl'. Did you called SNESSetFunction() to set your FormFunction1()?? On 10/3/07, JoaoFlavio Vasconcellos wrote: > Hello > > I'm trying to use SNES to solve a problem. I wrote the code above to evaluate the residual of the linear system. In this routine norm_2 of F is different from zero but outside of this routine norm_2 of F is zero. > > #undef __FUNCT__ > #define __FUNCT__ "FormFunction1" > PetscErrorCode FormFunction1(SNES snes, Vec X, Vec F, void *_simulator) { > Simulator *simulator = (Simulator*) _simulator; > PetscErrorCode ierr; > > VecView (X, 0); > ierr = simulator->CalculateMatrixCoefficient(X); CHKERRQ(ierr); > ierr = simulator->CalculateSourceTerm(X); CHKERRQ(ierr); > ierr = simulator->CalculateResidual (X, &F); CHKERRQ(ierr); > > cout << "Residual " << endl; > VecView (F, 0); // norm_2 of F is != zero > return 0; > > } > > // End of routine > > > > > Messages from -snes_view: > > 0 SNES Function norm 0.000000000000e+00 > SNES Object: > type: ls > line search variant: SNESLineSearchCubic > alpha=0.0001, maxstep=1e+08, steptol=1e-12 > maximum iterations=50, maximum function evaluations=10000 > tolerances: relative=1e-08, absolute=1e-50, solution=1e-08 > total number of linear solver iterations=0 > total number of function evaluations=1 > KSP Object: > type: gmres > GMRES: restart=30, using Classical (unmodified) Gram-Schmidt Orthogonalization with no iterative refinement > GMRES: happy breakdown tolerance 1e-30 > maximum iterations=10000, initial guess is zero > tolerances: relative=1e-05, absolute=1e-50, divergence=10000 > left preconditioning > PC Object: > type: ilu > ILU: 0 levels of fill > ILU: factor fill ratio allocated 1 > ILU: tolerance for zero pivot 1e-12 > out-of-place factorization > matrix ordering: natural > > // end of message > > I dont know why "SNES Function norm 0.000000000000e+00" if > F's norm_2 inside my routine is not. Has anyone that could sugest how to fix this problem ? > > > Thank you in advance. > > > Joao Flavio Vieira de Vasconcellos > School of Engineering > James Cook University > > Phone: +61 7 4781 4340 > Fax: +61 7 4775 1184 > Email: jflavio at iprj.uerj.br > joaoflavio.vasconcellos at jcu.edu.au > skype: jflavio_vasconcellos > msn: jf_vasconcellos at hotmail.com > > -- Lisandro Dalc?n --------------- Centro Internacional de M?todos Computacionales en Ingenier?a (CIMEC) Instituto de Desarrollo Tecnol?gico para la Industria Qu?mica (INTEC) Consejo Nacional de Investigaciones Cient?ficas y T?cnicas (CONICET) PTLC - G?emes 3450, (3000) Santa Fe, Argentina Tel/Fax: +54-(0)342-451.1594 From joaoflavio.vasconcellos at jcu.edu.au Wed Oct 3 16:32:14 2007 From: joaoflavio.vasconcellos at jcu.edu.au (JoaoFlavio Vasconcellos) Date: Thu, 4 Oct 2007 07:32:14 +1000 (EST) Subject: SNES problem Message-ID: <20071004073214.BMZ31320@mirapoint-ms1.jcu.edu.au> Dear Lisandro FormFunction1 is called only once. I'm quite sure that F is a non zero vector. I'd printed it and its values are 100% rigth. I've been used snes functions as written above: ... ... .... ierr = SNESCreate(PETSC_COMM_WORLD, &snes);CHKERRQ(ierr); ierr = SNESSetFunction(snes, resid, FormFunction1, this);CHKERRQ(ierr); .... .... // Calculating c_ctx.A .... ierr = FormJacobian1(snes, resid, &c_ctx.A, &c_ctx.A, &mat_str, this); CHKERRQ(ierr); ierr = SNESSetJacobian(snes, c_ctx.A, c_ctx.A, FormJacobian1, this); CHKERRQ(ierr); ierr = SNESSetFromOptions(snes);CHKERRQ(ierr); cout << "Calling SNESSolve" << endl; PetscScalar zero = 0; ierr = VecSet(C, zero);CHKERRQ(ierr); ierr = SNESSolve(snes,PETSC_NULL, C);CHKERRQ(ierr); ierr = SNESGetIterationNumber(snes, &its);CHKERRQ(ierr); ierr = SNESGetSolutionUpdate(snes, &C);CHKERRQ(ierr); VecView (C, 0); ... ... ... All outputs of VevView in FormFunction1 are ok. The output of VecView (C, 0); after call SNESSolve is an unexpected zero vector. Thanks JF Joao Flavio Vieira de Vasconcellos School of Engineering James Cook University Phone: +61 7 4781 4340 Fax: +61 7 4775 1184 Email: jflavio at iprj.uerj.br joaoflavio.vasconcellos at jcu.edu.au skype: jflavio_vasconcellos msn: jf_vasconcellos at hotmail.com ---- Original message ---- >Date: Wed, 3 Oct 2007 12:40:28 -0300 >From: "Lisandro Dalcin" >Subject: Re: SNES problem >To: petsc-users at mcs.anl.gov > >Flavio, are you completelly sure your FormFunction1() is actually >being called? I can see the SNES monitor at iter 0, but I did not see >in the the outputs of VecView() or your 'cout << "Residual " << endl'. >Did you called SNESSetFunction() to set your FormFunction1()?? > >On 10/3/07, JoaoFlavio Vasconcellos wrote: >> Hello >> >> I'm trying to use SNES to solve a problem. I wrote the code above to evaluate the residual of the linear system. In this routine norm_2 of F is different from zero but outside of this routine norm_2 of F is zero. >> >> #undef __FUNCT__ >> #define __FUNCT__ "FormFunction1" >> PetscErrorCode FormFunction1(SNES snes, Vec X, Vec F, void *_simulator) { >> Simulator *simulator = (Simulator*) _simulator; >> PetscErrorCode ierr; >> >> VecView (X, 0); >> ierr = simulator->CalculateMatrixCoefficient(X); CHKERRQ(ierr); >> ierr = simulator->CalculateSourceTerm(X); CHKERRQ(ierr); >> ierr = simulator->CalculateResidual (X, &F); CHKERRQ(ierr); >> >> cout << "Residual " << endl; >> VecView (F, 0); // norm_2 of F is != zero >> return 0; >> >> } >> >> // End of routine >> >> >> >> >> Messages from -snes_view: >> >> 0 SNES Function norm 0.000000000000e+00 >> SNES Object: >> type: ls >> line search variant: SNESLineSearchCubic >> alpha=0.0001, maxstep=1e+08, steptol=1e-12 >> maximum iterations=50, maximum function evaluations=10000 >> tolerances: relative=1e-08, absolute=1e-50, solution=1e-08 >> total number of linear solver iterations=0 >> total number of function evaluations=1 >> KSP Object: >> type: gmres >> GMRES: restart=30, using Classical (unmodified) Gram-Schmidt Orthogonalization with no iterative refinement >> GMRES: happy breakdown tolerance 1e-30 >> maximum iterations=10000, initial guess is zero >> tolerances: relative=1e-05, absolute=1e-50, divergence=10000 >> left preconditioning >> PC Object: >> type: ilu >> ILU: 0 levels of fill >> ILU: factor fill ratio allocated 1 >> ILU: tolerance for zero pivot 1e-12 >> out-of-place factorization >> matrix ordering: natural >> >> // end of message >> >> I dont know why "SNES Function norm 0.000000000000e+00" if >> F's norm_2 inside my routine is not. Has anyone that could sugest how to fix this problem ? >> >> >> Thank you in advance. >> >> >> Joao Flavio Vieira de Vasconcellos >> School of Engineering >> James Cook University >> >> Phone: +61 7 4781 4340 >> Fax: +61 7 4775 1184 >> Email: jflavio at iprj.uerj.br >> joaoflavio.vasconcellos at jcu.edu.au >> skype: jflavio_vasconcellos >> msn: jf_vasconcellos at hotmail.com >> >> > > >-- >Lisandro Dalc?n >--------------- >Centro Internacional de M?todos Computacionales en Ingenier?a (CIMEC) >Instituto de Desarrollo Tecnol?gico para la Industria Qu?mica (INTEC) >Consejo Nacional de Investigaciones Cient?ficas y T?cnicas (CONICET) >PTLC - G?emes 3450, (3000) Santa Fe, Argentina >Tel/Fax: +54-(0)342-451.1594 > From knepley at gmail.com Wed Oct 3 16:52:36 2007 From: knepley at gmail.com (Matthew Knepley) Date: Wed, 3 Oct 2007 16:52:36 -0500 Subject: SNES problem In-Reply-To: <20071004073214.BMZ31320@mirapoint-ms1.jcu.edu.au> References: <20071004073214.BMZ31320@mirapoint-ms1.jcu.edu.au> Message-ID: On 10/3/07, JoaoFlavio Vasconcellos wrote: > Dear Lisandro > > FormFunction1 is called only once. I'm quite sure that F is a non zero vector. I'd printed it and its values are 100% rigth. > > > I've been used snes functions as written above: > > ... > ... > .... > ierr = SNESCreate(PETSC_COMM_WORLD, &snes);CHKERRQ(ierr); > ierr = SNESSetFunction(snes, resid, FormFunction1, this);CHKERRQ(ierr); > .... > .... // Calculating c_ctx.A > .... > > ierr = FormJacobian1(snes, resid, &c_ctx.A, &c_ctx.A, &mat_str, this); CHKERRQ(ierr); > ierr = SNESSetJacobian(snes, c_ctx.A, c_ctx.A, FormJacobian1, this); CHKERRQ(ierr); > ierr = SNESSetFromOptions(snes);CHKERRQ(ierr); > cout << "Calling SNESSolve" << endl; > PetscScalar zero = 0; > ierr = VecSet(C, zero);CHKERRQ(ierr); > ierr = SNESSolve(snes,PETSC_NULL, C);CHKERRQ(ierr); > ierr = SNESGetIterationNumber(snes, &its);CHKERRQ(ierr); > ierr = SNESGetSolutionUpdate(snes, &C);CHKERRQ(ierr); > VecView (C, 0); > > > ... > ... > ... > > All outputs of VevView in FormFunction1 are ok. The output of VecView (C, 0); after call SNESSolve is an unexpected zero vector. It is hard to tell without seeing your code. However, there is something suspicious in your evaluation routine. You pass &F to CaculateResidual(). The 'Vec' type is just a pointer, which is passed to FormFunction() by value. You are supposed to use that Vec. If your routine changed the pointer, then FormFunction() would see the changed F and print a nonzero norm, but the calling routine would still have the original Vec which would be unchanged. Matt > Thanks > > JF > > Joao Flavio Vieira de Vasconcellos > School of Engineering > James Cook University > > Phone: +61 7 4781 4340 > Fax: +61 7 4775 1184 > Email: jflavio at iprj.uerj.br > joaoflavio.vasconcellos at jcu.edu.au > skype: jflavio_vasconcellos > msn: jf_vasconcellos at hotmail.com > > > ---- Original message ---- > >Date: Wed, 3 Oct 2007 12:40:28 -0300 > >From: "Lisandro Dalcin" > >Subject: Re: SNES problem > >To: petsc-users at mcs.anl.gov > > > >Flavio, are you completelly sure your FormFunction1() is actually > >being called? I can see the SNES monitor at iter 0, but I did not see > >in the the outputs of VecView() or your 'cout << "Residual " << endl'. > >Did you called SNESSetFunction() to set your FormFunction1()?? > > > >On 10/3/07, JoaoFlavio Vasconcellos wrote: > >> Hello > >> > >> I'm trying to use SNES to solve a problem. I wrote the code above to evaluate the residual of the linear system. In this routine norm_2 of F is different from zero but outside of this routine norm_2 of F is zero. > >> > >> #undef __FUNCT__ > >> #define __FUNCT__ "FormFunction1" > >> PetscErrorCode FormFunction1(SNES snes, Vec X, Vec F, void *_simulator) { > >> Simulator *simulator = (Simulator*) _simulator; > >> PetscErrorCode ierr; > >> > >> VecView (X, 0); > >> ierr = simulator->CalculateMatrixCoefficient(X); CHKERRQ(ierr); > >> ierr = simulator->CalculateSourceTerm(X); CHKERRQ(ierr); > >> ierr = simulator->CalculateResidual (X, &F); CHKERRQ(ierr); > >> > >> cout << "Residual " << endl; > >> VecView (F, 0); // norm_2 of F is != zero > >> return 0; > >> > >> } > >> > >> // End of routine > >> > >> > >> > >> > >> Messages from -snes_view: > >> > >> 0 SNES Function norm 0.000000000000e+00 > >> SNES Object: > >> type: ls > >> line search variant: SNESLineSearchCubic > >> alpha=0.0001, maxstep=1e+08, steptol=1e-12 > >> maximum iterations=50, maximum function evaluations=10000 > >> tolerances: relative=1e-08, absolute=1e-50, solution=1e-08 > >> total number of linear solver iterations=0 > >> total number of function evaluations=1 > >> KSP Object: > >> type: gmres > >> GMRES: restart=30, using Classical (unmodified) Gram-Schmidt Orthogonalization with no iterative refinement > >> GMRES: happy breakdown tolerance 1e-30 > >> maximum iterations=10000, initial guess is zero > >> tolerances: relative=1e-05, absolute=1e-50, divergence=10000 > >> left preconditioning > >> PC Object: > >> type: ilu > >> ILU: 0 levels of fill > >> ILU: factor fill ratio allocated 1 > >> ILU: tolerance for zero pivot 1e-12 > >> out-of-place factorization > >> matrix ordering: natural > >> > >> // end of message > >> > >> I dont know why "SNES Function norm 0.000000000000e+00" if > >> F's norm_2 inside my routine is not. Has anyone that could sugest how to fix this problem ? > >> > >> > >> Thank you in advance. > >> > >> > >> Joao Flavio Vieira de Vasconcellos > >> School of Engineering > >> James Cook University > >> > >> Phone: +61 7 4781 4340 > >> Fax: +61 7 4775 1184 > >> Email: jflavio at iprj.uerj.br > >> joaoflavio.vasconcellos at jcu.edu.au > >> skype: jflavio_vasconcellos > >> msn: jf_vasconcellos at hotmail.com > >> > >> > > > > > >-- > >Lisandro Dalc?n > >--------------- > >Centro Internacional de M?todos Computacionales en Ingenier?a (CIMEC) > >Instituto de Desarrollo Tecnol?gico para la Industria Qu?mica (INTEC) > >Consejo Nacional de Investigaciones Cient?ficas y T?cnicas (CONICET) > >PTLC - G?emes 3450, (3000) Santa Fe, Argentina > >Tel/Fax: +54-(0)342-451.1594 > > > > -- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. -- Norbert Wiener From joaoflavio.vasconcellos at jcu.edu.au Wed Oct 3 18:26:48 2007 From: joaoflavio.vasconcellos at jcu.edu.au (JoaoFlavio Vasconcellos) Date: Thu, 4 Oct 2007 09:26:48 +1000 (EST) Subject: SNES problem Message-ID: <20071004092648.BMZ35194@mirapoint-ms1.jcu.edu.au> Dear Matt, You were right. I'd made some modifications as you suggested and now the program is working. Thanks G'day Joao Flavio Vieira de Vasconcellos School of Engineering James Cook University Phone: +61 7 4781 4340 Fax: +61 7 4775 1184 Email: jflavio at iprj.uerj.br joaoflavio.vasconcellos at jcu.edu.au skype: jflavio_vasconcellos msn: jf_vasconcellos at hotmail.com ---- Original message ---- >Date: Wed, 3 Oct 2007 16:52:36 -0500 >From: "Matthew Knepley" >Subject: Re: SNES problem >To: petsc-users at users@users at users@mcs.anl.gov >Cc:"JoaoFlavioVasconcellos".vasconcellos at jcu.joaoflavio.vasconcellos@jcu.joaoflavio.vasconcellos at jcu.edu.au > >On10/3/07,JoaoFlavioVasconcellosjoaoflavio.vasconcellos at jcu.edu.auwrote >>DearLisandro >> >>FormFunction1iscalledonlyonce.I'mquitesurethatFisanonzerovector.I'dprinteditanditsvaluesare100%rigth. >> >> >>I'vebeenusedsnesfunctionsaswrittenabove: >> >> ... >> ... >> .... >> ierr = SNESCreate(PETSC_COMM_WORLD, &snes);CHKERRQ(ierr); >> ierr = SNESSetFunction(snes, resid, FormFunction1, this);CHKERRQ(ierr); >> .... >> .... // Calculating c_ctx.A >> .... >> >> ierr = FormJacobian1(snes, resid, &c_ctx.A, &c_ctx.A, &mat_str, this); CHKERRQ(ierr); >> ierr = SNESSetJacobian(snes, c_ctx.A, c_ctx.A, FormJacobian1, this); CHKERRQ(ierr); >> ierr = SNESSetFromOptions(snes);CHKERRQ(ierr); >> cout << "Calling SNESSolve" << endl; >> PetscScalar zero = 0; >> ierr = VecSet(C, zero);CHKERRQ(ierr); >> ierr = SNESSolve(snes,PETSC_NULL, C);CHKERRQ(ierr); >> ierr = SNESGetIterationNumber(snes, &its);CHKERRQ(ierr); >> ierr = SNESGetSolutionUpdate(snes, &C);CHKERRQ(ierr); >> VecView (C, 0); >> >> >> ... >> ... >> ... >> >> All outputs of VevView in FormFunction1 are ok. The output of VecView (C, 0); after call SNESSolve is an unexpected zero vector. > >It is hard to tell without seeing your code. However, there is >something suspicious >in your evaluation routine. You pass &F to CaculateResidual(). The 'Vec' type is >just a pointer, which is passed to FormFunction() by value. You are supposed to >use that Vec. If your routine changed the pointer, then FormFunction() would see >the changed F and print a nonzero norm, but the calling routine would still have >the original Vec which would be unchanged. > > Matt > >> Thanks >> >> JF >> >> Joao Flavio Vieira de Vasconcellos >> School of Engineering >> James Cook University >> >> Phone: +61 7 4781 4340 >> Fax: +61 7 4775 1184 >> Email: jflavio at iprj.uerj.br >> joaoflavio.vasconcellos at jcu.edu.au >> skype: jflavio_vasconcellos >> msn: jf_vasconcellos at hotmail.com >> >> >> ---- Original message ---- >> >Date: Wed, 3 Oct 2007 12:40:28 -0300 >> >From: "Lisandro Dalcin" >> >Subject: Re: SNES problem >> >To: petsc-users at users@users at users@mcs.anl.gov >>> >>>Flavio,areyoucompletellysureyourFormFunction1()isactually >>>being called? I can see the SNES monitor at iter 0, but I did not see >> >in the the outputs of VecView() or your 'cout << "Residual " << endl'. >> >Did you called SNESSetFunction() to set your FormFunction1()?? >> > >> >On 10/3/07, JoaoFlavio Vasconcellos wrote: >> >> Hello >> >> >> >> I'm trying to use SNES to solve a problem. I wrote the code above to evaluate the residual of the linear system. In this routine norm_2 of F is different from zero but outside of this routine norm_2 of F is zero. >> >> >> >> #undef __FUNCT__ >> >> #define __FUNCT__ "FormFunction1" >> >> PetscErrorCode FormFunction1(SNES snes, Vec X, Vec F, void *_simulator) { >> >> Simulator *simulator = (Simulator*) _simulator; >> >> PetscErrorCode ierr; >> >> >> >> VecView (X, 0); >> >> ierr = simulator->CalculateMatrixCoefficient(X); CHKERRQ(ierr); >> >> ierr = simulator->CalculateSourceTerm(X); CHKERRQ(ierr); >> >> ierr = simulator->CalculateResidual (X, &F); CHKERRQ(ierr); >> >> >> >> cout << "Residual " << endl; >> >> VecView (F, 0); // norm_2 of F is != zero >> >> return 0; >> >> >> >> } >> >> >> >> // End of routine >> >> >> >> >> >> >> >> >> >> Messages from -snes_view: >> >> >> >> 0 SNES Function norm 0.000000000000e+00 >> >> SNES Object: >> >> type: ls >> >> line search variant: SNESLineSearchCubic >> >> alpha=0.0001, maxstep=1e+08, steptol=1e-12 >> >> maximum iterations=50, maximum function evaluations=10000 >> >> tolerances: relative=1e-08, absolute=1e-50, solution=1e-08 >> >> total number of linear solver iterations=0 >> >> total number of function evaluations=1 >> >> KSP Object: >> >> type: gmres >> >> GMRES: restart=30, using Classical (unmodified) Gram-Schmidt Orthogonalization with no iterative refinement >> >> GMRES: happy breakdown tolerance 1e-30 >> >> maximum iterations=10000, initial guess is zero >> >> tolerances: relative=1e-05, absolute=1e-50, divergence=10000 >> >> left preconditioning >> >> PC Object: >> >> type: ilu >> >> ILU: 0 levels of fill >> >> ILU: factor fill ratio allocated 1 >> >> ILU: tolerance for zero pivot 1e-12 >> >> out-of-place factorization >> >> matrix ordering: natural >> >> >> >> // end of message >> >> >> >> I dont know why "SNES Function norm 0.000000000000e+00" if >> >> F's norm_2 inside my routine is not. Has anyone that could sugest how to fix this problem ? >> >> >> >> >> >> Thank you in advance. >> >> >> >> >> >> Joao Flavio Vieira de Vasconcellos >> >> School of Engineering >> >> James Cook University >> >> >> >> Phone: +61 7 4781 4340 >> >> Fax: +61 7 4775 1184 >> >> Email: jflavio at iprj.uerj.br >> >> joaoflavio.vasconcellos at jcu.edu.au >> >> skype: jflavio_vasconcellos >> >> msn: jf_vasconcellos at hotmail.com >> >> >> >> >> > >> > >> >-- >> >Lisandro Dalc?n >> >--------------- >> >Centro Internacional de M?todos Computacionales en Ingenier?a (CIMEC) >> >Instituto de Desarrollo Tecnol?gico para la Industria Qu?mica (INTEC) >> >Consejo Nacional de Investigaciones Cient?ficas y T?cnicas (CONICET) >> >PTLC - G?emes 3450, (3000) Santa Fe, Argentina >> >Tel/Fax: +54-(0)342-451.1594 >> > >> >> > > >-- >What most experimenters take for granted before they begin their >experiments is infinitely more interesting than any results to which >their experiments lead. >-- Norbert Wiener > From timothy.stitt at ichec.ie Thu Oct 4 06:29:11 2007 From: timothy.stitt at ichec.ie (Dr. Timothy Stitt) Date: Thu, 04 Oct 2007 12:29:11 +0100 Subject: Order (N) Sparse Banded Linear Solver Message-ID: <4704CE87.70202@ichec.ie> Dear PETSc Community, I was wondering if anyone knew of any sparse/dense banded linear solvers that run in order (N)? Does PETSc provide such a solver or link to an appropriate external solver? There seems to be a lot of links online to theoretical approaches but we are having difficulties in sourcing an implementation? I am hoping PETSc might be our saviour ??? Thanks in advance, Tim. -- Dr. Timothy Stitt HPC Application Consultant - ICHEC (www.ichec.ie) Dublin Institute for Advanced Studies 5 Merrion Square - Dublin 2 - Ireland +353-1-6621333 (tel) / +353-1-6621477 (fax) / +353-874195427 (mobile) From knepley at gmail.com Thu Oct 4 07:15:57 2007 From: knepley at gmail.com (Matthew Knepley) Date: Thu, 4 Oct 2007 07:15:57 -0500 Subject: Order (N) Sparse Banded Linear Solver In-Reply-To: <4704CE87.70202@ichec.ie> References: <4704CE87.70202@ichec.ie> Message-ID: On 10/4/07, Dr. Timothy Stitt wrote: > Dear PETSc Community, > > I was wondering if anyone knew of any sparse/dense banded linear solvers > that run in order (N)? Does PETSc provide such a solver or link to an > appropriate external solver? There seems to be a lot of links online to > theoretical approaches but we are having difficulties in sourcing an > implementation? I am hoping PETSc might be our saviour ??? Have you tried MUMPS or SuperLU, both available through PETSc's configure? They are not truly O(N), but its hard to tell the difference until you get very very large, or have a nice elliptic PDE on a regular grid. In the latter case, you should use the PETSc DMMG. Matt > Thanks in advance, > > Tim. > > -- > Dr. Timothy Stitt > HPC Application Consultant - ICHEC (www.ichec.ie) > > Dublin Institute for Advanced Studies > 5 Merrion Square - Dublin 2 - Ireland > > +353-1-6621333 (tel) / +353-1-6621477 (fax) / +353-874195427 (mobile) > > -- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. -- Norbert Wiener From timothy.stitt at ichec.ie Thu Oct 4 08:55:25 2007 From: timothy.stitt at ichec.ie (Dr. Timothy Stitt) Date: Thu, 04 Oct 2007 14:55:25 +0100 Subject: Order (N) Sparse Banded Linear Solver In-Reply-To: References: <4704CE87.70202@ichec.ie> Message-ID: <4704F0CD.1000803@ichec.ie> Matthew, Out of curiosity what would you term as very very large? The matrices I am using will probably be of order 10^3 x 10^3 elements. Thanks, Tim. Matthew Knepley wrote: > On 10/4/07, Dr. Timothy Stitt wrote: > >> Dear PETSc Community, >> >> I was wondering if anyone knew of any sparse/dense banded linear solvers >> that run in order (N)? Does PETSc provide such a solver or link to an >> appropriate external solver? There seems to be a lot of links online to >> theoretical approaches but we are having difficulties in sourcing an >> implementation? I am hoping PETSc might be our saviour ??? >> > > Have you tried MUMPS or SuperLU, both available through PETSc's configure? > They are not truly O(N), but its hard to tell the difference until you get very > very large, or have a nice elliptic PDE on a regular grid. In the > latter case, you > should use the PETSc DMMG. > > Matt > > >> Thanks in advance, >> >> Tim. >> >> -- >> Dr. Timothy Stitt >> HPC Application Consultant - ICHEC (www.ichec.ie) >> >> Dublin Institute for Advanced Studies >> 5 Merrion Square - Dublin 2 - Ireland >> >> +353-1-6621333 (tel) / +353-1-6621477 (fax) / +353-874195427 (mobile) >> >> >> > > > -- Dr. Timothy Stitt HPC Application Consultant - ICHEC (www.ichec.ie) Dublin Institute for Advanced Studies 5 Merrion Square - Dublin 2 - Ireland +353-1-6621333 (tel) / +353-1-6621477 (fax) / +353-874195427 (mobile) From knepley at gmail.com Thu Oct 4 09:01:54 2007 From: knepley at gmail.com (Matthew Knepley) Date: Thu, 4 Oct 2007 09:01:54 -0500 Subject: Order (N) Sparse Banded Linear Solver In-Reply-To: <4704F0CD.1000803@ichec.ie> References: <4704CE87.70202@ichec.ie> <4704F0CD.1000803@ichec.ie> Message-ID: On 10/4/07, Dr. Timothy Stitt wrote: > Matthew, > > Out of curiosity what would you term as very very large? The matrices I > am using will probably be of order 10^3 x 10^3 elements. I believe this is fairly small. SuperLU and MUMPS routinely report results for matrices of order 10^5x10^5 (on large parallel systems). Matt > Thanks, > > Tim. > > Matthew Knepley wrote: > > On 10/4/07, Dr. Timothy Stitt wrote: > > > >> Dear PETSc Community, > >> > >> I was wondering if anyone knew of any sparse/dense banded linear solvers > >> that run in order (N)? Does PETSc provide such a solver or link to an > >> appropriate external solver? There seems to be a lot of links online to > >> theoretical approaches but we are having difficulties in sourcing an > >> implementation? I am hoping PETSc might be our saviour ??? > >> > > > > Have you tried MUMPS or SuperLU, both available through PETSc's configure? > > They are not truly O(N), but its hard to tell the difference until you get very > > very large, or have a nice elliptic PDE on a regular grid. In the > > latter case, you > > should use the PETSc DMMG. > > > > Matt > > > > > >> Thanks in advance, > >> > >> Tim. > >> > >> -- > >> Dr. Timothy Stitt > >> HPC Application Consultant - ICHEC (www.ichec.ie) > >> > >> Dublin Institute for Advanced Studies > >> 5 Merrion Square - Dublin 2 - Ireland > >> > >> +353-1-6621333 (tel) / +353-1-6621477 (fax) / +353-874195427 (mobile) > >> > >> > >> > > > > > > > > > -- > Dr. Timothy Stitt > HPC Application Consultant - ICHEC (www.ichec.ie) > > Dublin Institute for Advanced Studies > 5 Merrion Square - Dublin 2 - Ireland > > +353-1-6621333 (tel) / +353-1-6621477 (fax) / +353-874195427 (mobile) > > -- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. -- Norbert Wiener From bsmith at mcs.anl.gov Thu Oct 4 10:36:40 2007 From: bsmith at mcs.anl.gov (Barry Smith) Date: Thu, 4 Oct 2007 10:36:40 -0500 (CDT) Subject: Compiler Optimization Override In-Reply-To: <200706271208.44921.timothy.stitt@ichec.ie> References: <200706271208.44921.timothy.stitt@ichec.ie> Message-ID: It seems the config/configure.py options -COPTFLAGS, -CXXOPTFLAGS, -FOPTFLAGS work correctly. That is if you set them then the default optimization is turned off and what you provide is used. This works with petsc-dev. If this does not work for someone please send configure.log so we can resolve the problem. Barry On Wed, 27 Jun 2007, Tim Stitt wrote: > Hi PETSc Users/Developers, > > Could someone explain how I can override the default optimization level that > comes with my PETSc build (currently -O). I want to set a more aggressive > optimization level (e.g. -O3) for my compiler (PathScale). > > Ideally I would prefer a solution which doesn't involve rebuilding PETSc each > time I want to switch the optimization level. I tried to add the switch to > the petscrules file but it always defaulted to the first (lower?) > optimization switch that appeared in the application compile command line > which seems to be built into PETSc. > > Any help gratefully received. > > Regards, > > Tim. > > From berend at chalmers.se Thu Oct 4 11:02:44 2007 From: berend at chalmers.se (Berend van Wachem) Date: Thu, 04 Oct 2007 18:02:44 +0200 Subject: Compiler Optimization Override In-Reply-To: References: <200706271208.44921.timothy.stitt@ichec.ie> Message-ID: <47050EA4.3020701@chalmers.se> Hi Tim, >> Ideally I would prefer a solution which doesn't involve rebuilding PETSc each >> time I want to switch the optimization level. I tried to add the switch to >> the petscrules file but it always defaulted to the first (lower?) >> optimization switch that appeared in the application compile command line >> which seems to be built into PETSc. I do this by having a different PETSC_ARCH setting for each different configuration. Then you only have to change this variable and rebuild your application. Berend. From recrusader at gmail.com Sun Oct 7 01:27:17 2007 From: recrusader at gmail.com (Yujie) Date: Sun, 7 Oct 2007 14:27:17 +0800 Subject: which solving methods is more efficient to my problem. Message-ID: <7ff0ee010710062327i583253a8j3fd1bc6b95f62424@mail.gmail.com> Hi, everyone I am using adaptive mesh refinement to solve a group of coupled steady-state linear PDEs. I will use PETSC as the solver. Did anyone have experience in such problem. I want to know which solving method is more efficient, including iterative method and preconditioner. Thanks a lot. Regards, Yujie -------------- next part -------------- An HTML attachment was scrubbed... URL: From renzhengyong at gmail.com Sun Oct 7 07:30:51 2007 From: renzhengyong at gmail.com (RenZhengYong) Date: Sun, 7 Oct 2007 20:30:51 +0800 Subject: which solving methods is more efficient to my problem. In-Reply-To: <7ff0ee010710062327i583253a8j3fd1bc6b95f62424@mail.gmail.com> References: <7ff0ee010710062327i583253a8j3fd1bc6b95f62424@mail.gmail.com> Message-ID: <4913f8f50710070530u5e37fe5aoe82db8cb0c51e567@mail.gmail.com> I have did some similar job. Based on my results, the AMG method is optimal. On 10/7/07, Yujie wrote: > > Hi, everyone > > I am using adaptive mesh refinement to solve a group of coupled > steady-state linear PDEs. I will use PETSC as the solver. > Did anyone have experience in such problem. I want to know which solving method is more efficient, including iterative method and preconditioner. Thanks a lot. > > > Regards, > Yujie > -- Zhengyong Ren School of Info-physics and Geomatics Engineering DiXue Building, 3-322#, Central South University, Hunan Province,China. Postal code: 410083 Email: renzhengyong at gmail.com Tel: +086013787102372. -------------- next part -------------- An HTML attachment was scrubbed... URL: From kamaraju at bluebottle.com Sun Oct 7 17:05:02 2007 From: kamaraju at bluebottle.com (Kamaraju Kusumanchi) Date: Sun, 7 Oct 2007 18:05:02 -0400 Subject: migrating from IMSL to petsc Message-ID: <200710071805.02534.kamaraju@bluebottle.com> Hi all I have a rather large Fortran 90 code (say 10,000 lines long) which extensively uses IMSL libraries for solving linear systems. Because of the limited capabilities of IMSL, the cost involved in purchasing the license etc., we are considering going with petsc instead. I have read the FAQs on the petsc website, finished reading the user manual. I have a few questions which the experts might find trivial to answer... Let's say I am solving Ax=B using IMSL's LU decomposition routine. In this case, A would be an nxn matrix, x and B would nx1 vectors. In the fortran 90 code these are declared as (assuming statically allocated) real :: A(1:n,1:n), B(1:n), x(1:n) The IMSL routines just directly accept A, x, B as arguments. Whereas, the petsc's solvers seem to require the arguments to be of type Vec, Mat etc., Converting all the arrays into Vec, Mat types would involve quite a bit of changes in the code, which I would like to avoid if possible. So my questions are 1) What is the best way to migrate a fortran 90 code using IMSL libraries to Petsc libraries with minimal changes in the code? 2) Is there any way to use petsc linear solvers (to solve Ax=b) without dealing with new data types Vec, Mat etc., and just sticking with real arrays? At this point, I am only interested in solving the system of equations Ax=b and am not trying to exploit other cool features of petsc. thanks for any suggestions raju -- Kamaraju S Kusumanchi http://www.people.cornell.edu/pages/kk288/ http://malayamaarutham.blogspot.com/ ---------------------------------------------------------------------- Find out how you can get spam free email. http://www.bluebottle.com/tag/3 From knutert at stud.ntnu.no Mon Oct 8 02:38:03 2007 From: knutert at stud.ntnu.no (knutert at stud.ntnu.no) Date: Mon, 8 Oct 2007 09:38:03 +0200 Subject: migrating from IMSL to petsc In-Reply-To: <200710071805.02534.kamaraju@bluebottle.com> References: <200710071805.02534.kamaraju@bluebottle.com> Message-ID: <20071008093803.c65t3od4e88884wo@webmail.ntnu.no> Hello, I have converted a code of similar size to use PETSc solvers. What I did was to write small, generic subroutines to convert arrays to Vec and Mat format. Very small changes to the actual code had to be made, and the conversion process went relatively painless. Good luck, maybe the PETSc developers have some better suggestions. -Knut Erik- Siterer Kamaraju Kusumanchi : > Hi all > > I have a rather large Fortran 90 code (say 10,000 lines long) which > extensively uses IMSL libraries for solving linear systems. Because of the > limited capabilities of IMSL, the cost involved in purchasing the license > etc., we are considering going with petsc instead. I have read the FAQs on > the petsc website, finished reading the user manual. I have a few questions > which the experts might find trivial to answer... > > Let's say I am solving Ax=B using IMSL's LU decomposition > routine. In this > case, A would be an nxn matrix, x and B would nx1 vectors. In the fortran 90 > code these are declared as (assuming statically allocated) > > real :: A(1:n,1:n), B(1:n), x(1:n) > > The IMSL routines just directly accept A, x, B as arguments. Whereas, the > petsc's solvers seem to require the arguments to be of type Vec, Mat etc., > Converting all the arrays into Vec, Mat types would involve quite a bit of > changes in the code, which I would like to avoid if possible. So my questions > are > > 1) What is the best way to migrate a fortran 90 code using IMSL libraries to > Petsc libraries with minimal changes in the code? > > 2) Is there any way to use petsc linear solvers (to solve Ax=b) without > dealing with new data types Vec, Mat etc., and just sticking with real > arrays? At this point, I am only interested in solving the system of > equations Ax=b and am not trying to exploit other cool features of petsc. > > thanks for any suggestions > raju > > -- > Kamaraju S Kusumanchi > http://www.people.cornell.edu/pages/kk288/ > http://malayamaarutham.blogspot.com/ > > ---------------------------------------------------------------------- > Find out how you can get spam free email. > http://www.bluebottle.com/tag/3 > > > From tim.sears at anu.edu.au Mon Oct 8 03:50:23 2007 From: tim.sears at anu.edu.au (Tim Sears) Date: Mon, 8 Oct 2007 18:50:23 +1000 Subject: Block sparse matrix with repeated values Message-ID: I wish to perform the matrix-vector multiplication Ax, where A = (I 'kron' B), and B is a rectangular matrix. 'kron' is the Kronecker product of two matrices. Clearly A contains many repeated values. Is there a PETSC format that can store this matrix efficiently? If so how does one populate it? Cheers, --Tim From knepley at gmail.com Mon Oct 8 04:32:20 2007 From: knepley at gmail.com (Matthew Knepley) Date: Mon, 8 Oct 2007 04:32:20 -0500 Subject: Block sparse matrix with repeated values In-Reply-To: References: Message-ID: We have no format that handles Kronecker products explicitly. However, these generally reduce to one-dimensional versions, and are maintained unassembled I think. However, maybe Barry has a better suggestion. Matt On 10/8/07, Tim Sears wrote: > I wish to perform the matrix-vector multiplication Ax, where A = (I > 'kron' B), and B is a rectangular matrix. > 'kron' is the Kronecker product of two matrices. > Clearly A contains many repeated values. Is there a PETSC format that > can store this matrix efficiently? > If so how does one populate it? > Cheers, > --Tim > > > -- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. -- Norbert Wiener From tim.sears at anu.edu.au Mon Oct 8 04:49:20 2007 From: tim.sears at anu.edu.au (Tim Sears) Date: Mon, 8 Oct 2007 19:49:20 +1000 Subject: Block sparse matrix with repeated values In-Reply-To: References: Message-ID: The kron here merely involves the identity matrix, so any solution that promotes reusing the memory holding B would be useful. Hopefully ending with a call to MatMult. Thanks, Tim On Oct 8, 2007, at 7:32 PM, Matthew Knepley wrote: > We have no format that handles Kronecker products explicitly. However, > these generally reduce to one-dimensional versions, and are maintained > unassembled I think. However, maybe Barry has a better suggestion. > > Matt > > On 10/8/07, Tim Sears wrote: >> I wish to perform the matrix-vector multiplication Ax, where A = (I >> 'kron' B), and B is a rectangular matrix. >> 'kron' is the Kronecker product of two matrices. >> Clearly A contains many repeated values. Is there a PETSC format that >> can store this matrix efficiently? >> If so how does one populate it? >> Cheers, >> --Tim >> >> >> > > > -- > What most experimenters take for granted before they begin their > experiments is infinitely more interesting than any results to which > their experiments lead. > -- Norbert Wiener > From knepley at gmail.com Mon Oct 8 08:17:29 2007 From: knepley at gmail.com (Matthew Knepley) Date: Mon, 8 Oct 2007 08:17:29 -0500 Subject: migrating from IMSL to petsc In-Reply-To: <200710071805.02534.kamaraju@bluebottle.com> References: <200710071805.02534.kamaraju@bluebottle.com> Message-ID: On 10/7/07, Kamaraju Kusumanchi wrote: > Hi all > > I have a rather large Fortran 90 code (say 10,000 lines long) which > extensively uses IMSL libraries for solving linear systems. Because of the > limited capabilities of IMSL, the cost involved in purchasing the license > etc., we are considering going with petsc instead. I have read the FAQs on > the petsc website, finished reading the user manual. I have a few questions > which the experts might find trivial to answer... > > Let's say I am solving Ax=B using IMSL's LU decomposition routine. In this > case, A would be an nxn matrix, x and B would nx1 vectors. In the fortran 90 > code these are declared as (assuming statically allocated) > > real :: A(1:n,1:n), B(1:n), x(1:n) > > The IMSL routines just directly accept A, x, B as arguments. Whereas, the > petsc's solvers seem to require the arguments to be of type Vec, Mat etc., > Converting all the arrays into Vec, Mat types would involve quite a bit of > changes in the code, which I would like to avoid if possible. So my questions > are > > 1) What is the best way to migrate a fortran 90 code using IMSL libraries to > Petsc libraries with minimal changes in the code? a) For vectors, the situation is very good. You can pull out the vector as an F90 array using VecGetArrayF90(). If you really want to manage your own memory, we can easily create a VecCreateWithArrayF90() for you, however I really suggest the first solution since it will work correctly in parallel. b) For matrices, it is slightly for complicated. If you stick to your current strategy, you can use MatGetArrayF90(). However, this fails to generalize for other matrix storage formats. Thus, I generally suggest moving to the MatSetValues() paradigm, but the user is in charge. > 2) Is there any way to use petsc linear solvers (to solve Ax=b) without > dealing with new data types Vec, Mat etc., and just sticking with real > arrays? At this point, I am only interested in solving the system of > equations Ax=b and am not trying to exploit other cool features of petsc. No, but as I pointed out above. You should not have to change your code in order to use Mat and Vec. Matt > thanks for any suggestions > raju > > -- > Kamaraju S Kusumanchi > http://www.people.cornell.edu/pages/kk288/ > http://malayamaarutham.blogspot.com/ > > ---------------------------------------------------------------------- > Find out how you can get spam free email. > http://www.bluebottle.com/tag/3 > > -- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. -- Norbert Wiener From dalcinl at gmail.com Mon Oct 8 10:56:09 2007 From: dalcinl at gmail.com (Lisandro Dalcin) Date: Mon, 8 Oct 2007 12:56:09 -0300 Subject: which solving methods is more efficient to my problem. In-Reply-To: <7ff0ee010710062327i583253a8j3fd1bc6b95f62424@mail.gmail.com> References: <7ff0ee010710062327i583253a8j3fd1bc6b95f62424@mail.gmail.com> Message-ID: You forgot to mention what kind of problem your PDE is, appart from being linear. I mean, can you describe which kind of problem are you trying to solve? On 10/7/07, Yujie wrote: > Hi, everyone > > I am using adaptive mesh refinement to solve a group of coupled steady-state > linear PDEs. I will use PETSC as the solver. Did anyone have experience > in such problem. I want to know which > solving method is more efficient, including > iterative method and preconditioner. > Thanks a lot. > > Regards, > Yujie > -- Lisandro Dalc?n --------------- Centro Internacional de M?todos Computacionales en Ingenier?a (CIMEC) Instituto de Desarrollo Tecnol?gico para la Industria Qu?mica (INTEC) Consejo Nacional de Investigaciones Cient?ficas y T?cnicas (CONICET) PTLC - G?emes 3450, (3000) Santa Fe, Argentina Tel/Fax: +54-(0)342-451.1594 From bsmith at mcs.anl.gov Mon Oct 8 12:10:05 2007 From: bsmith at mcs.anl.gov (Barry Smith) Date: Mon, 8 Oct 2007 12:10:05 -0500 (CDT) Subject: migrating from IMSL to petsc In-Reply-To: <200710071805.02534.kamaraju@bluebottle.com> References: <200710071805.02534.kamaraju@bluebottle.com> Message-ID: Use Vec vb,vx Mat mA call VecCreateSeqWithArray(PETSC_COMM_SELF,n,b,vb,ierr) call VecCreateSeqWithArray(PETSC_COMM_SELF,n,x,vx,ierr) call MatCreateSeqDense(PETSC_COMM_SELF,n,n,A,mA,ierr) call KSPCreate(PETSC_COMM_SELF,ksp,ierr) call KSPSetType(ksp,KSPPREONLY,ierr) call KSPGetPC(ksp,pc,ierr) call PCSetType(pc,PCLU,ierr) call KSPSetOperators(ksp,mA,mA,SAME_NONZERO_PATTERN,ierr) call KSPSolve(ksp,vb,vx,ierr) call KSPDestroy(ksp,ierr) call MatDestroy(mA,ierr) call VecDestroy(vb,ierr) call VecDestroy(vx,ierr) The VecCreateSeqWithArray(), MatCreateSeqDense() also have parallel versions with Seq replaced with MPI. Barry On Sun, 7 Oct 2007, Kamaraju Kusumanchi wrote: > Hi all > > I have a rather large Fortran 90 code (say 10,000 lines long) which > extensively uses IMSL libraries for solving linear systems. Because of the > limited capabilities of IMSL, the cost involved in purchasing the license > etc., we are considering going with petsc instead. I have read the FAQs on > the petsc website, finished reading the user manual. I have a few questions > which the experts might find trivial to answer... > > Let's say I am solving Ax=B using IMSL's LU decomposition routine. In this > case, A would be an nxn matrix, x and B would nx1 vectors. In the fortran 90 > code these are declared as (assuming statically allocated) > > real :: A(1:n,1:n), B(1:n), x(1:n) > > The IMSL routines just directly accept A, x, B as arguments. Whereas, the > petsc's solvers seem to require the arguments to be of type Vec, Mat etc., > Converting all the arrays into Vec, Mat types would involve quite a bit of > changes in the code, which I would like to avoid if possible. So my questions > are > > 1) What is the best way to migrate a fortran 90 code using IMSL libraries to > Petsc libraries with minimal changes in the code? > > 2) Is there any way to use petsc linear solvers (to solve Ax=b) without > dealing with new data types Vec, Mat etc., and just sticking with real > arrays? At this point, I am only interested in solving the system of > equations Ax=b and am not trying to exploit other cool features of petsc. > > thanks for any suggestions > raju > > From bsmith at mcs.anl.gov Mon Oct 8 12:23:25 2007 From: bsmith at mcs.anl.gov (Barry Smith) Date: Mon, 8 Oct 2007 12:23:25 -0500 (CDT) Subject: Block sparse matrix with repeated values In-Reply-To: References: Message-ID: Unfortunately we do not have code specifically for this purpose. One would need to write a MATSHELL matrix that then uses repeated multiplications against another Mat (representing the B presumably). Barry On Mon, 8 Oct 2007, Matthew Knepley wrote: > We have no format that handles Kronecker products explicitly. However, > these generally reduce to one-dimensional versions, and are maintained > unassembled I think. However, maybe Barry has a better suggestion. > > Matt > > On 10/8/07, Tim Sears wrote: > > I wish to perform the matrix-vector multiplication Ax, where A = (I > > 'kron' B), and B is a rectangular matrix. > > 'kron' is the Kronecker product of two matrices. > > Clearly A contains many repeated values. Is there a PETSC format that > > can store this matrix efficiently? > > If so how does one populate it? > > Cheers, > > --Tim > > > > > > > > > From berend at chalmers.se Tue Oct 9 07:59:29 2007 From: berend at chalmers.se (Berend van Wachem) Date: Tue, 09 Oct 2007 14:59:29 +0200 Subject: Petsc Error Message-ID: <470B7B31.9070803@chalmers.se> Hi, On the command ierr = DACreateLocalVector(da,g); CHKERRQ(ierr); I get the error [0]PETSC ERROR: PetscMallocAlign() line 61 in src/sys/memory/mal.c Memory requested 6050600 [0]PETSC ERROR: PETSC: Attaching gdb to Debug/MultiFlow of pid 19087 on display :0.0 on machine tfdpc102 What does it mean exactly? Out of memory? Thanks, Berend. From knepley at gmail.com Tue Oct 9 08:13:31 2007 From: knepley at gmail.com (Matthew Knepley) Date: Tue, 9 Oct 2007 08:13:31 -0500 Subject: Petsc Error In-Reply-To: <470B7B31.9070803@chalmers.se> References: <470B7B31.9070803@chalmers.se> Message-ID: On 10/9/07, Berend van Wachem wrote: > Hi, > > On the command > > ierr = DACreateLocalVector(da,g); CHKERRQ(ierr); > > I get the error > > > [0]PETSC ERROR: PetscMallocAlign() line 61 in src/sys/memory/mal.c > Memory requested 6050600 > [0]PETSC ERROR: PETSC: Attaching gdb to Debug/MultiFlow of pid 19087 on > display :0.0 on machine tfdpc102 > > > What does it mean exactly? Out of memory? Yes. This is strange. The line above these two should read "Out of memory". Matt > Thanks, > Berend. > > -- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. -- Norbert Wiener From berend at chalmers.se Tue Oct 9 10:25:04 2007 From: berend at chalmers.se (Berend van Wachem) Date: Tue, 09 Oct 2007 17:25:04 +0200 Subject: Petsc Error In-Reply-To: References: <470B7B31.9070803@chalmers.se> Message-ID: <470B9D50.8060901@chalmers.se> Hi, I ran the same command again, running top, and got the result Mem: 6220940k total, 5879056k used, 341884k free, 143992k buffers Swap: 1951888k total, 2584k used, 1949304k free, 2363516k cached PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 19117 berend 25 0 3037m 2.5g 3508 R 100 41.4 0:44.96 MultiFlow just before it crashed. I have a 2 processor machine with total 6 GB of memory (2 times 3 gb?), so I don't think it should be out of memory. How can I verify this for sure? Thanks, Berend. Matthew Knepley wrote: > On 10/9/07, Berend van Wachem wrote: >> Hi, >> >> On the command >> >> ierr = DACreateLocalVector(da,g); CHKERRQ(ierr); >> >> I get the error >> >> >> [0]PETSC ERROR: PetscMallocAlign() line 61 in src/sys/memory/mal.c >> Memory requested 6050600 >> [0]PETSC ERROR: PETSC: Attaching gdb to Debug/MultiFlow of pid 19087 on >> display :0.0 on machine tfdpc102 >> >> >> What does it mean exactly? Out of memory? > > Yes. This is strange. The line above these two should read "Out of memory". > > Matt > >> Thanks, >> Berend. >> >> > > From knepley at gmail.com Tue Oct 9 10:28:57 2007 From: knepley at gmail.com (Matthew Knepley) Date: Tue, 9 Oct 2007 10:28:57 -0500 Subject: Petsc Error In-Reply-To: <470B9D50.8060901@chalmers.se> References: <470B7B31.9070803@chalmers.se> <470B9D50.8060901@chalmers.se> Message-ID: On 10/9/07, Berend van Wachem wrote: > Hi, > > I ran the same command again, running top, and got the result > > Mem: 6220940k total, 5879056k used, 341884k free, 143992k buffers > Swap: 1951888k total, 2584k used, 1949304k free, 2363516k cached > > PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND > 19117 berend 25 0 3037m 2.5g 3508 R 100 41.4 0:44.96 MultiFlow > > > > just before it crashed. I have a 2 processor machine with total 6 GB of > memory (2 times 3 gb?), so I don't think it should be out of memory. How > can I verify this for sure? Run in the debugger and break on mal.c:61. malloc() is returning a null, but you can look at errno to see if something crazy is happening. Matt > Thanks, > > Berend. > > > > Matthew Knepley wrote: > > On 10/9/07, Berend van Wachem wrote: > >> Hi, > >> > >> On the command > >> > >> ierr = DACreateLocalVector(da,g); CHKERRQ(ierr); > >> > >> I get the error > >> > >> > >> [0]PETSC ERROR: PetscMallocAlign() line 61 in src/sys/memory/mal.c > >> Memory requested 6050600 > >> [0]PETSC ERROR: PETSC: Attaching gdb to Debug/MultiFlow of pid 19087 on > >> display :0.0 on machine tfdpc102 > >> > >> > >> What does it mean exactly? Out of memory? > > > > Yes. This is strange. The line above these two should read "Out of memory". > > > > Matt > > > >> Thanks, > >> Berend. > >> > >> > > > > > > -- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. -- Norbert Wiener From bsmith at mcs.anl.gov Tue Oct 9 11:59:32 2007 From: bsmith at mcs.anl.gov (Barry Smith) Date: Tue, 9 Oct 2007 11:59:32 -0500 (CDT) Subject: Petsc Error In-Reply-To: <470B9D50.8060901@chalmers.se> References: <470B7B31.9070803@chalmers.se> <470B9D50.8060901@chalmers.se> Message-ID: On most OS's just because you have greater than a certain amount of physical memory does not mean that a SINGLE process can utilize it. For example, many times a single process is limited to 2, 3 or 4 gigabytes. This is EVEN with 64 bit pointers (and certainly with 32 bit). Barry Run a simple C program that mallocs a chunk of memory. How much you you malloc? On Tue, 9 Oct 2007, Berend van Wachem wrote: > Hi, > > I ran the same command again, running top, and got the result > > Mem: 6220940k total, 5879056k used, 341884k free, 143992k buffers > Swap: 1951888k total, 2584k used, 1949304k free, 2363516k cached > > PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND > 19117 berend 25 0 3037m 2.5g 3508 R 100 41.4 0:44.96 MultiFlow > > > > just before it crashed. I have a 2 processor machine with total 6 GB of memory > (2 times 3 gb?), so I don't think it should be out of memory. How can I verify > this for sure? > > Thanks, > > Berend. > > > > Matthew Knepley wrote: > > On 10/9/07, Berend van Wachem wrote: > > > Hi, > > > > > > On the command > > > > > > ierr = DACreateLocalVector(da,g); CHKERRQ(ierr); > > > > > > I get the error > > > > > > > > > [0]PETSC ERROR: PetscMallocAlign() line 61 in src/sys/memory/mal.c > > > Memory requested 6050600 > > > [0]PETSC ERROR: PETSC: Attaching gdb to Debug/MultiFlow of pid 19087 on > > > display :0.0 on machine tfdpc102 > > > > > > > > > What does it mean exactly? Out of memory? > > > > Yes. This is strange. The line above these two should read "Out of memory". > > > > Matt > > > > > Thanks, > > > Berend. > > > > > > > > > > > > From berend at chalmers.se Tue Oct 9 12:34:22 2007 From: berend at chalmers.se (Berend van Wachem) Date: Tue, 09 Oct 2007 19:34:22 +0200 Subject: Petsc Error In-Reply-To: References: <470B7B31.9070803@chalmers.se> <470B9D50.8060901@chalmers.se> Message-ID: <470BBB9E.7060505@chalmers.se> Hi Barry, Yes, your hint seems correct; a single process can "only" allocate 3 gigabytes of memory, even if there is more physical memory present. Is the only work around to run parallel? Thanks, Berend. Barry Smith wrote: > On most OS's just because you have greater than a certain amount of > physical memory does not mean that a SINGLE process can utilize it. > For example, many times a single process is limited to 2, 3 or 4 gigabytes. > This is EVEN with 64 bit pointers (and certainly with 32 bit). > > Barry > > Run a simple C program that mallocs a chunk of memory. How much you > you malloc? > > On Tue, 9 Oct 2007, Berend van Wachem wrote: > >> Hi, >> >> I ran the same command again, running top, and got the result >> >> Mem: 6220940k total, 5879056k used, 341884k free, 143992k buffers >> Swap: 1951888k total, 2584k used, 1949304k free, 2363516k cached >> >> PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND >> 19117 berend 25 0 3037m 2.5g 3508 R 100 41.4 0:44.96 MultiFlow >> >> >> >> just before it crashed. I have a 2 processor machine with total 6 GB of memory >> (2 times 3 gb?), so I don't think it should be out of memory. How can I verify >> this for sure? >> >> Thanks, >> >> Berend. >> >> >> >> Matthew Knepley wrote: >>> On 10/9/07, Berend van Wachem wrote: >>>> Hi, >>>> >>>> On the command >>>> >>>> ierr = DACreateLocalVector(da,g); CHKERRQ(ierr); >>>> >>>> I get the error >>>> >>>> >>>> [0]PETSC ERROR: PetscMallocAlign() line 61 in src/sys/memory/mal.c >>>> Memory requested 6050600 >>>> [0]PETSC ERROR: PETSC: Attaching gdb to Debug/MultiFlow of pid 19087 on >>>> display :0.0 on machine tfdpc102 >>>> >>>> >>>> What does it mean exactly? Out of memory? >>> Yes. This is strange. The line above these two should read "Out of memory". >>> >>> Matt >>> >>>> Thanks, >>>> Berend. >>>> >>>> >>> >> > From bsmith at mcs.anl.gov Tue Oct 9 12:40:05 2007 From: bsmith at mcs.anl.gov (Barry Smith) Date: Tue, 9 Oct 2007 12:40:05 -0500 (CDT) Subject: Petsc Error In-Reply-To: <470BBB9E.7060505@chalmers.se> References: <470B7B31.9070803@chalmers.se> <470B9D50.8060901@chalmers.se> <470BBB9E.7060505@chalmers.se> Message-ID: Pressure your vendors, be they IBM or Redhat, to provide a 2007 OS instead of a 2001 OS :-(. Barry On Tue, 9 Oct 2007, Berend van Wachem wrote: > Hi Barry, > > Yes, your hint seems correct; a single process can "only" allocate 3 gigabytes > of memory, even if there is more physical memory present. > > Is the only work around to run parallel? > > Thanks, > Berend. > > > Barry Smith wrote: > > On most OS's just because you have greater than a certain amount of > > physical memory does not mean that a SINGLE process can utilize it. > > For example, many times a single process is limited to 2, 3 or 4 gigabytes. > > This is EVEN with 64 bit pointers (and certainly with 32 bit). > > > > Barry > > > > Run a simple C program that mallocs a chunk of memory. How much you you > > malloc? > > > > On Tue, 9 Oct 2007, Berend van Wachem wrote: > > > > > Hi, > > > > > > I ran the same command again, running top, and got the result > > > > > > Mem: 6220940k total, 5879056k used, 341884k free, 143992k buffers > > > Swap: 1951888k total, 2584k used, 1949304k free, 2363516k cached > > > > > > PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND > > > 19117 berend 25 0 3037m 2.5g 3508 R 100 41.4 0:44.96 MultiFlow > > > > > > > > > > > > just before it crashed. I have a 2 processor machine with total 6 GB of > > > memory > > > (2 times 3 gb?), so I don't think it should be out of memory. How can I > > > verify > > > this for sure? > > > > > > Thanks, > > > > > > Berend. > > > > > > > > > > > > Matthew Knepley wrote: > > > > On 10/9/07, Berend van Wachem wrote: > > > > > Hi, > > > > > > > > > > On the command > > > > > > > > > > ierr = DACreateLocalVector(da,g); CHKERRQ(ierr); > > > > > > > > > > I get the error > > > > > > > > > > > > > > > [0]PETSC ERROR: PetscMallocAlign() line 61 in src/sys/memory/mal.c > > > > > Memory requested 6050600 > > > > > [0]PETSC ERROR: PETSC: Attaching gdb to Debug/MultiFlow of pid 19087 > > > > > on > > > > > display :0.0 on machine tfdpc102 > > > > > > > > > > > > > > > What does it mean exactly? Out of memory? > > > > Yes. This is strange. The line above these two should read "Out of > > > > memory". > > > > > > > > Matt > > > > > > > > > Thanks, > > > > > Berend. > > > > > > > > > > > > > > > > > > > > > From berend at chalmers.se Tue Oct 9 12:43:00 2007 From: berend at chalmers.se (Berend van Wachem) Date: Tue, 09 Oct 2007 19:43:00 +0200 Subject: Petsc Error In-Reply-To: References: <470B7B31.9070803@chalmers.se> <470B9D50.8060901@chalmers.se> <470BBB9E.7060505@chalmers.se> Message-ID: <470BBDA4.4060404@chalmers.se> Hi, I am using 32 bit linux (slackware) on a 64 bit opteron. Would using a 64 bit linux distro help? Berend. Barry Smith wrote: > Pressure your vendors, be they IBM or Redhat, to provide a 2007 OS instead > of a 2001 OS :-(. > > Barry > > > On Tue, 9 Oct 2007, Berend van Wachem wrote: > >> Hi Barry, >> >> Yes, your hint seems correct; a single process can "only" allocate 3 gigabytes >> of memory, even if there is more physical memory present. >> >> Is the only work around to run parallel? >> >> Thanks, >> Berend. >> >> >> Barry Smith wrote: >>> On most OS's just because you have greater than a certain amount of >>> physical memory does not mean that a SINGLE process can utilize it. >>> For example, many times a single process is limited to 2, 3 or 4 gigabytes. >>> This is EVEN with 64 bit pointers (and certainly with 32 bit). >>> >>> Barry >>> >>> Run a simple C program that mallocs a chunk of memory. How much you you >>> malloc? >>> >>> On Tue, 9 Oct 2007, Berend van Wachem wrote: >>> >>>> Hi, >>>> >>>> I ran the same command again, running top, and got the result >>>> >>>> Mem: 6220940k total, 5879056k used, 341884k free, 143992k buffers >>>> Swap: 1951888k total, 2584k used, 1949304k free, 2363516k cached >>>> >>>> PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND >>>> 19117 berend 25 0 3037m 2.5g 3508 R 100 41.4 0:44.96 MultiFlow >>>> >>>> >>>> >>>> just before it crashed. I have a 2 processor machine with total 6 GB of >>>> memory >>>> (2 times 3 gb?), so I don't think it should be out of memory. How can I >>>> verify >>>> this for sure? >>>> >>>> Thanks, >>>> >>>> Berend. >>>> >>>> >>>> >>>> Matthew Knepley wrote: >>>>> On 10/9/07, Berend van Wachem wrote: >>>>>> Hi, >>>>>> >>>>>> On the command >>>>>> >>>>>> ierr = DACreateLocalVector(da,g); CHKERRQ(ierr); >>>>>> >>>>>> I get the error >>>>>> >>>>>> >>>>>> [0]PETSC ERROR: PetscMallocAlign() line 61 in src/sys/memory/mal.c >>>>>> Memory requested 6050600 >>>>>> [0]PETSC ERROR: PETSC: Attaching gdb to Debug/MultiFlow of pid 19087 >>>>>> on >>>>>> display :0.0 on machine tfdpc102 >>>>>> >>>>>> >>>>>> What does it mean exactly? Out of memory? >>>>> Yes. This is strange. The line above these two should read "Out of >>>>> memory". >>>>> >>>>> Matt >>>>> >>>>>> Thanks, >>>>>> Berend. >>>>>> >>>>>> >> > From balay at mcs.anl.gov Tue Oct 9 12:54:19 2007 From: balay at mcs.anl.gov (Satish Balay) Date: Tue, 9 Oct 2007 12:54:19 -0500 (CDT) Subject: Petsc Error In-Reply-To: <470BBDA4.4060404@chalmers.se> References: <470B7B31.9070803@chalmers.se> <470B9D50.8060901@chalmers.se> <470BBB9E.7060505@chalmers.se> <470BBDA4.4060404@chalmers.se> Message-ID: On Tue, 9 Oct 2007, Berend van Wachem wrote: > Hi, > > I am using 32 bit linux (slackware) on a 64 bit opteron. Would using a 64 bit > linux distro help? 32bit adressing => 4GB max. [both for the OS and a process]. But generally a single process allocate can allocate only 2GB or 3GB - not 4GB. [This is OS dependent - as the OS requires the process to use this remaining addressspace for other things]. 64bit linux will avoid this limitation. Satish From balay at mcs.anl.gov Tue Oct 9 13:09:19 2007 From: balay at mcs.anl.gov (Satish Balay) Date: Tue, 9 Oct 2007 13:09:19 -0500 (CDT) Subject: Petsc Error In-Reply-To: References: <470B7B31.9070803@chalmers.se> <470B9D50.8060901@chalmers.se> <470BBB9E.7060505@chalmers.se> <470BBDA4.4060404@chalmers.se> Message-ID: On Tue, 9 Oct 2007, Satish Balay wrote: > On Tue, 9 Oct 2007, Berend van Wachem wrote: > > > Hi, > > > > I am using 32 bit linux (slackware) on a 64 bit opteron. Would using a 64 bit > > linux distro help? > > 32bit adressing => 4GB max. [both for the OS and a process]. But > generally a single process allocate can allocate only 2GB or 3GB - not > 4GB. [This is OS dependent - as the OS requires the process to use > this remaining addressspace for other things]. > > 64bit linux will avoid this limitation. Another note: Looks like your current 32bit OS is able to see all the 6GB of memory. This is possible with bigmem kernel - which uses 36 address bits to manage this. http://en.wikipedia.org/wiki/Physical_Address_Extension However each process still uses 32bit adressing - and its memory usage limited to 2-3GB. Satish From jagruti.trivedi at navy.mil Tue Oct 9 13:21:08 2007 From: jagruti.trivedi at navy.mil (Trivedi, Jagruti CIV 470000D, 474300D) Date: Tue, 9 Oct 2007 11:21:08 -0700 Subject: Compiled information for Petsc 2.3.2-p10 Message-ID: Hello there, I am trying to port my application from petsc 2.2.1 to petsc 2.3.2-p10. Has anybody compiled The real life changes needed to make other than listed in the changes file? I really appreciate if you have hands on do's and don'ts for porting from petsc-2.2.1 to Petsc 2.3.2-p10. Looking forward to positive replies. Thanks and Best Regards, ....Jagruti NAWC -------------- next part -------------- An HTML attachment was scrubbed... URL: From berend at chalmers.se Tue Oct 9 13:25:24 2007 From: berend at chalmers.se (Berend van Wachem) Date: Tue, 09 Oct 2007 20:25:24 +0200 Subject: Compiled information for Petsc 2.3.2-p10 In-Reply-To: References: Message-ID: <470BC794.8060109@chalmers.se> Hi, I'm not sure if it is optimal, but I usually go through the changes when upgrading, http://www-unix.mcs.anl.gov/petsc/petsc-as/documentation/changes/index.html Maybe there is a more "easy" way, I'm not sure. Berend. Trivedi, Jagruti CIV 470000D, 474300D wrote: > > > Hello there, > > I am trying to port my application from petsc 2.2.1 to petsc 2.3.2-p10. > Has anybody compiled > The real life changes needed to make other than listed in the changes file? > > I really appreciate if you have hands on do's and don'ts for porting > from petsc-2.2.1 to Petsc 2.3.2-p10. > > Looking forward to positive replies. > > Thanks and Best Regards, > > ?.Jagruti > > NAWC > From balay at mcs.anl.gov Tue Oct 9 13:28:48 2007 From: balay at mcs.anl.gov (Satish Balay) Date: Tue, 9 Oct 2007 13:28:48 -0500 (CDT) Subject: Compiled information for Petsc 2.3.2-p10 In-Reply-To: References: Message-ID: It is possible that we didn't document all the changes between releases. But the recommended approach is to look at changes file for each version after 2.2.1 - and apply these changes. And if your code is in C - then the compiler will help in finding any undocumented changes in prototypes.. etc.. BTW: the current release version is 2.3.3 [so you might want to upgrade to this version instead of 2.3.2] Satish On Tue, 9 Oct 2007, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > > Hello there, > > I am trying to port my application from petsc 2.2.1 to petsc 2.3.2-p10. > Has anybody compiled > The real life changes needed to make other than listed in the changes > file? > > I really appreciate if you have hands on do's and don'ts for porting > from petsc-2.2.1 to Petsc 2.3.2-p10. > > Looking forward to positive replies. > > Thanks and Best Regards, > > ....Jagruti > > NAWC > From jagruti.trivedi at navy.mil Tue Oct 9 14:19:05 2007 From: jagruti.trivedi at navy.mil (Trivedi, Jagruti CIV 470000D, 474300D) Date: Tue, 9 Oct 2007 12:19:05 -0700 Subject: Compiled information for Petsc 2.3.2-p10 In-Reply-To: References: Message-ID: Do you mean that I should apply changes from 2.2.1 to 2.3.0 and then 2.3.1 and so forth until I come upto 2.3.3 version or I can apply directly from 2.2.1 to 2.3.3? ...Jagruti -----Original Message----- From: owner-petsc-users at mcs.anl.gov [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Satish Balay Sent: Tuesday, October 09, 2007 11:29 To: petsc-users at mcs.anl.gov Subject: Re: Compiled information for Petsc 2.3.2-p10 It is possible that we didn't document all the changes between releases. But the recommended approach is to look at changes file for each version after 2.2.1 - and apply these changes. And if your code is in C - then the compiler will help in finding any undocumented changes in prototypes.. etc.. BTW: the current release version is 2.3.3 [so you might want to upgrade to this version instead of 2.3.2] Satish On Tue, 9 Oct 2007, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > > Hello there, > > I am trying to port my application from petsc 2.2.1 to petsc 2.3.2-p10. > Has anybody compiled > The real life changes needed to make other than listed in the changes > file? > > I really appreciate if you have hands on do's and don'ts for porting > from petsc-2.2.1 to Petsc 2.3.2-p10. > > Looking forward to positive replies. > > Thanks and Best Regards, > > ....Jagruti > > NAWC > From dalcinl at gmail.com Tue Oct 9 14:30:37 2007 From: dalcinl at gmail.com (Lisandro Dalcin) Date: Tue, 9 Oct 2007 16:30:37 -0300 Subject: Compiled information for Petsc 2.3.2-p10 In-Reply-To: References: Message-ID: I want to suggest you to update to 2.3.3. Of course, you have the extra trouble of figuring out yourself the cummulative changes since 2.2.1. Go ahead updating to 2.3.3, and I can help you at any point you have trouble. On 10/9/07, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > Do you mean that I should apply changes from 2.2.1 to 2.3.0 and then > 2.3.1 and so forth until I come upto 2.3.3 version or I can apply > directly from 2.2.1 to 2.3.3? > > ...Jagruti > -----Original Message----- > From: owner-petsc-users at mcs.anl.gov > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Satish Balay > Sent: Tuesday, October 09, 2007 11:29 > To: petsc-users at mcs.anl.gov > Subject: Re: Compiled information for Petsc 2.3.2-p10 > > > It is possible that we didn't document all the changes between releases. > > But the recommended approach is to look at changes file for each version > after 2.2.1 - and apply these changes. > > And if your code is in C - then the compiler will help in finding any > undocumented changes in prototypes.. etc.. > > BTW: the current release version is 2.3.3 [so you might want to upgrade > to this version instead of 2.3.2] > > Satish > > On Tue, 9 Oct 2007, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > > > > > > Hello there, > > > > I am trying to port my application from petsc 2.2.1 to petsc > 2.3.2-p10. > > Has anybody compiled > > The real life changes needed to make other than listed in the changes > > file? > > > > I really appreciate if you have hands on do's and don'ts for porting > > from petsc-2.2.1 to Petsc 2.3.2-p10. > > > > Looking forward to positive replies. > > > > Thanks and Best Regards, > > > > ....Jagruti > > > > NAWC > > > > -- Lisandro Dalc?n --------------- Centro Internacional de M?todos Computacionales en Ingenier?a (CIMEC) Instituto de Desarrollo Tecnol?gico para la Industria Qu?mica (INTEC) Consejo Nacional de Investigaciones Cient?ficas y T?cnicas (CONICET) PTLC - G?emes 3450, (3000) Santa Fe, Argentina Tel/Fax: +54-(0)342-451.1594 From jagruti.trivedi at navy.mil Tue Oct 9 17:45:26 2007 From: jagruti.trivedi at navy.mil (Trivedi, Jagruti CIV 470000D, 474300D) Date: Tue, 9 Oct 2007 15:45:26 -0700 Subject: Compiled information for Petsc 2.3.2-p10 In-Reply-To: References: Message-ID: Hello Licendro, Here is the first question after the installation of petsc-2.3.3-p6. I ran: Config/config.py --with-cc=gcc --with-fc=g77 --download-f-blas-lapack=1 --download-mpich=1 /usr/local/bin/make all test /* run gnu make I get the following status error: Running test examples to verify correct installation c/C++ example src/snes/examples/tutorials/ex19 run successfully with 1 MPI process c/C++ example src/snes/examples/tutorials/ex19 run successfully with 2 MPI process Graphics example src/snes/examples/tutorials/ex19 run successfully with 1 MPI process /bin/sh: test: argument expected make[3]: [ex19.rm] Error 1 (ignored) /bin/sh: test: argument expected make[3]: [ex19.rm] Error 1 (ignored) /bin/sh: test: argument expected make[3]: [ex19.rm] Error 1 (ignored) /bin/sh: test: argument expected make[3]: [ex19.rm] Error 1 (ignored) Fortran example src/snes/examples/tutorials/ex5f run successfully with 1 MPI process /bin/sh: test: argument expected make[3]: [ex5f.rm] Error 1 (ignored) /bin/sh: test: argument expected make[3]: [ex5f.rm] Error 1 (ignored) /bin/sh: test: argument expected make[3]: [ex5f.rm] Error 1 (ignored) /bin/sh: test: argument expected make[3]: [ex5f.rm] Error 1 (ignored) Completed test examples Is this the correct installation? Why am I getting this error messages? Thanks, ....Jagruti -----Original Message----- From: owner-petsc-users at mcs.anl.gov [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Lisandro Dalcin Sent: Tuesday, October 09, 2007 12:31 To: petsc-users at mcs.anl.gov Subject: Re: Compiled information for Petsc 2.3.2-p10 I want to suggest you to update to 2.3.3. Of course, you have the extra trouble of figuring out yourself the cummulative changes since 2.2.1. Go ahead updating to 2.3.3, and I can help you at any point you have trouble. On 10/9/07, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > Do you mean that I should apply changes from 2.2.1 to 2.3.0 and then > 2.3.1 and so forth until I come upto 2.3.3 version or I can apply > directly from 2.2.1 to 2.3.3? > > ...Jagruti > -----Original Message----- > From: owner-petsc-users at mcs.anl.gov > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Satish Balay > Sent: Tuesday, October 09, 2007 11:29 > To: petsc-users at mcs.anl.gov > Subject: Re: Compiled information for Petsc 2.3.2-p10 > > > It is possible that we didn't document all the changes between releases. > > But the recommended approach is to look at changes file for each > version after 2.2.1 - and apply these changes. > > And if your code is in C - then the compiler will help in finding any > undocumented changes in prototypes.. etc.. > > BTW: the current release version is 2.3.3 [so you might want to > upgrade to this version instead of 2.3.2] > > Satish > > On Tue, 9 Oct 2007, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > > > > > > Hello there, > > > > I am trying to port my application from petsc 2.2.1 to petsc > 2.3.2-p10. > > Has anybody compiled > > The real life changes needed to make other than listed in the > > changes file? > > > > I really appreciate if you have hands on do's and don'ts for porting > > from petsc-2.2.1 to Petsc 2.3.2-p10. > > > > Looking forward to positive replies. > > > > Thanks and Best Regards, > > > > ....Jagruti > > > > NAWC > > > > -- Lisandro Dalc?n --------------- Centro Internacional de M?todos Computacionales en Ingenier?a (CIMEC) Instituto de Desarrollo Tecnol?gico para la Industria Qu?mica (INTEC) Consejo Nacional de Investigaciones Cient?ficas y T?cnicas (CONICET) PTLC - G?emes 3450, (3000) Santa Fe, Argentina Tel/Fax: +54-(0)342-451.1594 From bsmith at mcs.anl.gov Tue Oct 9 18:47:23 2007 From: bsmith at mcs.anl.gov (Barry Smith) Date: Tue, 9 Oct 2007 18:47:23 -0500 (CDT) Subject: Compiled information for Petsc 2.3.2-p10 In-Reply-To: References: Message-ID: Strange, never seen that before. It is likely you can just ignore it. Barry On Tue, 9 Oct 2007, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > Hello Licendro, > > Here is the first question after the installation of petsc-2.3.3-p6. > > I ran: > Config/config.py --with-cc=gcc --with-fc=g77 --download-f-blas-lapack=1 --download-mpich=1 > > /usr/local/bin/make all test /* run gnu make > > I get the following status error: > > Running test examples to verify correct installation > c/C++ example src/snes/examples/tutorials/ex19 run successfully with 1 MPI process > > c/C++ example src/snes/examples/tutorials/ex19 run successfully with 2 MPI process > > Graphics example src/snes/examples/tutorials/ex19 run successfully with 1 MPI process > > /bin/sh: test: argument expected > make[3]: [ex19.rm] Error 1 (ignored) > /bin/sh: test: argument expected > make[3]: [ex19.rm] Error 1 (ignored) > /bin/sh: test: argument expected > make[3]: [ex19.rm] Error 1 (ignored) > /bin/sh: test: argument expected > make[3]: [ex19.rm] Error 1 (ignored) > Fortran example src/snes/examples/tutorials/ex5f run successfully with 1 MPI process > /bin/sh: test: argument expected > make[3]: [ex5f.rm] Error 1 (ignored) > /bin/sh: test: argument expected > make[3]: [ex5f.rm] Error 1 (ignored) > /bin/sh: test: argument expected > make[3]: [ex5f.rm] Error 1 (ignored) > /bin/sh: test: argument expected > make[3]: [ex5f.rm] Error 1 (ignored) > Completed test examples > > Is this the correct installation? Why am I getting this error messages? > > Thanks, > > ....Jagruti > > -----Original Message----- > From: owner-petsc-users at mcs.anl.gov [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Lisandro Dalcin > Sent: Tuesday, October 09, 2007 12:31 > To: petsc-users at mcs.anl.gov > Subject: Re: Compiled information for Petsc 2.3.2-p10 > > I want to suggest you to update to 2.3.3. Of course, you have the extra trouble of figuring out yourself the cummulative changes since 2.2.1. > > Go ahead updating to 2.3.3, and I can help you at any point you have trouble. > > > > On 10/9/07, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > > > Do you mean that I should apply changes from 2.2.1 to 2.3.0 and then > > 2.3.1 and so forth until I come upto 2.3.3 version or I can apply > > directly from 2.2.1 to 2.3.3? > > > > ...Jagruti > > -----Original Message----- > > From: owner-petsc-users at mcs.anl.gov > > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Satish Balay > > Sent: Tuesday, October 09, 2007 11:29 > > To: petsc-users at mcs.anl.gov > > Subject: Re: Compiled information for Petsc 2.3.2-p10 > > > > > > It is possible that we didn't document all the changes between releases. > > > > But the recommended approach is to look at changes file for each > > version after 2.2.1 - and apply these changes. > > > > And if your code is in C - then the compiler will help in finding any > > undocumented changes in prototypes.. etc.. > > > > BTW: the current release version is 2.3.3 [so you might want to > > upgrade to this version instead of 2.3.2] > > > > Satish > > > > On Tue, 9 Oct 2007, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > > > > > > > > > > Hello there, > > > > > > I am trying to port my application from petsc 2.2.1 to petsc > > 2.3.2-p10. > > > Has anybody compiled > > > The real life changes needed to make other than listed in the > > > changes file? > > > > > > I really appreciate if you have hands on do's and don'ts for porting > > > from petsc-2.2.1 to Petsc 2.3.2-p10. > > > > > > Looking forward to positive replies. > > > > > > Thanks and Best Regards, > > > > > > ....Jagruti > > > > > > NAWC > > > > > > > > > > -- > Lisandro Dalc?n > --------------- > Centro Internacional de M?todos Computacionales en Ingenier?a (CIMEC) Instituto de Desarrollo Tecnol?gico para la Industria Qu?mica (INTEC) Consejo Nacional de Investigaciones Cient?ficas y T?cnicas (CONICET) PTLC - G?emes 3450, (3000) Santa Fe, Argentina > Tel/Fax: +54-(0)342-451.1594 > > From dalcinl at gmail.com Tue Oct 9 19:08:13 2007 From: dalcinl at gmail.com (Lisandro Dalcin) Date: Tue, 9 Oct 2007 21:08:13 -0300 Subject: Compiled information for Petsc 2.3.2-p10 In-Reply-To: References: Message-ID: Perhaps PETSc is not internally calling '/usr/local/bin/make' ?? Was '/usr/local/bin/test' in front of your $PATH when you configured PETSc? As Barry said, you can just ignore this, unless you have problems building your own source. In that case, please send 'configure.log' file. On 10/9/07, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > Hello Licendro, > > Here is the first question after the installation of petsc-2.3.3-p6. > > I ran: > Config/config.py --with-cc=gcc --with-fc=g77 --download-f-blas-lapack=1 --download-mpich=1 > > /usr/local/bin/make all test /* run gnu make > > I get the following status error: > > Running test examples to verify correct installation > c/C++ example src/snes/examples/tutorials/ex19 run successfully with 1 MPI process > > c/C++ example src/snes/examples/tutorials/ex19 run successfully with 2 MPI process > > Graphics example src/snes/examples/tutorials/ex19 run successfully with 1 MPI process > > /bin/sh: test: argument expected > make[3]: [ex19.rm] Error 1 (ignored) > /bin/sh: test: argument expected > make[3]: [ex19.rm] Error 1 (ignored) > /bin/sh: test: argument expected > make[3]: [ex19.rm] Error 1 (ignored) > /bin/sh: test: argument expected > make[3]: [ex19.rm] Error 1 (ignored) > Fortran example src/snes/examples/tutorials/ex5f run successfully with 1 MPI process > /bin/sh: test: argument expected > make[3]: [ex5f.rm] Error 1 (ignored) > /bin/sh: test: argument expected > make[3]: [ex5f.rm] Error 1 (ignored) > /bin/sh: test: argument expected > make[3]: [ex5f.rm] Error 1 (ignored) > /bin/sh: test: argument expected > make[3]: [ex5f.rm] Error 1 (ignored) > Completed test examples > > Is this the correct installation? Why am I getting this error messages? > > Thanks, > > ....Jagruti > > -----Original Message----- > From: owner-petsc-users at mcs.anl.gov [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Lisandro Dalcin > Sent: Tuesday, October 09, 2007 12:31 > To: petsc-users at mcs.anl.gov > Subject: Re: Compiled information for Petsc 2.3.2-p10 > > I want to suggest you to update to 2.3.3. Of course, you have the extra trouble of figuring out yourself the cummulative changes since 2.2.1. > > Go ahead updating to 2.3.3, and I can help you at any point you have trouble. > > > > On 10/9/07, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > > > Do you mean that I should apply changes from 2.2.1 to 2.3.0 and then > > 2.3.1 and so forth until I come upto 2.3.3 version or I can apply > > directly from 2.2.1 to 2.3.3? > > > > ...Jagruti > > -----Original Message----- > > From: owner-petsc-users at mcs.anl.gov > > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Satish Balay > > Sent: Tuesday, October 09, 2007 11:29 > > To: petsc-users at mcs.anl.gov > > Subject: Re: Compiled information for Petsc 2.3.2-p10 > > > > > > It is possible that we didn't document all the changes between releases. > > > > But the recommended approach is to look at changes file for each > > version after 2.2.1 - and apply these changes. > > > > And if your code is in C - then the compiler will help in finding any > > undocumented changes in prototypes.. etc.. > > > > BTW: the current release version is 2.3.3 [so you might want to > > upgrade to this version instead of 2.3.2] > > > > Satish > > > > On Tue, 9 Oct 2007, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > > > > > > > > > > Hello there, > > > > > > I am trying to port my application from petsc 2.2.1 to petsc > > 2.3.2-p10. > > > Has anybody compiled > > > The real life changes needed to make other than listed in the > > > changes file? > > > > > > I really appreciate if you have hands on do's and don'ts for porting > > > from petsc-2.2.1 to Petsc 2.3.2-p10. > > > > > > Looking forward to positive replies. > > > > > > Thanks and Best Regards, > > > > > > ....Jagruti > > > > > > NAWC > > > > > > > > > > -- > Lisandro Dalc?n > --------------- > Centro Internacional de M?todos Computacionales en Ingenier?a (CIMEC) Instituto de Desarrollo Tecnol?gico para la Industria Qu?mica (INTEC) Consejo Nacional de Investigaciones Cient?ficas y T?cnicas (CONICET) PTLC - G?emes 3450, (3000) Santa Fe, Argentina > Tel/Fax: +54-(0)342-451.1594 > > -- Lisandro Dalc?n --------------- Centro Internacional de M?todos Computacionales en Ingenier?a (CIMEC) Instituto de Desarrollo Tecnol?gico para la Industria Qu?mica (INTEC) Consejo Nacional de Investigaciones Cient?ficas y T?cnicas (CONICET) PTLC - G?emes 3450, (3000) Santa Fe, Argentina Tel/Fax: +54-(0)342-451.1594 From jagruti.trivedi at navy.mil Tue Oct 9 19:14:34 2007 From: jagruti.trivedi at navy.mil (Trivedi, Jagruti CIV 470000D, 474300D) Date: Tue, 9 Oct 2007 17:14:34 -0700 Subject: Compiled information for Petsc 2.3.2-p10 In-Reply-To: References: Message-ID: Barry, When I compile my application it fails with a linking "ld" error that it can not Find -lmpich. With this new install, I have mpich2-1.0.5p4. Where do I change from mpich to mpich2 linking? Thanks, ...Jagruti -----Original Message----- From: owner-petsc-users at mcs.anl.gov [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Barry Smith Sent: Tuesday, October 09, 2007 16:47 To: petsc-users at mcs.anl.gov Subject: RE: Compiled information for Petsc 2.3.2-p10 Strange, never seen that before. It is likely you can just ignore it. Barry On Tue, 9 Oct 2007, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > Hello Licendro, > > Here is the first question after the installation of petsc-2.3.3-p6. > > I ran: > Config/config.py --with-cc=gcc --with-fc=g77 > --download-f-blas-lapack=1 --download-mpich=1 > > /usr/local/bin/make all test /* run gnu make > > I get the following status error: > > Running test examples to verify correct installation c/C++ example > src/snes/examples/tutorials/ex19 run successfully with 1 MPI process > > c/C++ example src/snes/examples/tutorials/ex19 run successfully with 2 > MPI process > > Graphics example src/snes/examples/tutorials/ex19 run successfully > with 1 MPI process > > /bin/sh: test: argument expected > make[3]: [ex19.rm] Error 1 (ignored) > /bin/sh: test: argument expected > make[3]: [ex19.rm] Error 1 (ignored) > /bin/sh: test: argument expected > make[3]: [ex19.rm] Error 1 (ignored) > /bin/sh: test: argument expected > make[3]: [ex19.rm] Error 1 (ignored) > Fortran example src/snes/examples/tutorials/ex5f run successfully with > 1 MPI process > /bin/sh: test: argument expected > make[3]: [ex5f.rm] Error 1 (ignored) > /bin/sh: test: argument expected > make[3]: [ex5f.rm] Error 1 (ignored) > /bin/sh: test: argument expected > make[3]: [ex5f.rm] Error 1 (ignored) > /bin/sh: test: argument expected > make[3]: [ex5f.rm] Error 1 (ignored) > Completed test examples > > Is this the correct installation? Why am I getting this error messages? > > Thanks, > > ....Jagruti > > -----Original Message----- > From: owner-petsc-users at mcs.anl.gov > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Lisandro Dalcin > Sent: Tuesday, October 09, 2007 12:31 > To: petsc-users at mcs.anl.gov > Subject: Re: Compiled information for Petsc 2.3.2-p10 > > I want to suggest you to update to 2.3.3. Of course, you have the extra trouble of figuring out yourself the cummulative changes since 2.2.1. > > Go ahead updating to 2.3.3, and I can help you at any point you have trouble. > > > > On 10/9/07, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > > > Do you mean that I should apply changes from 2.2.1 to 2.3.0 and then > > 2.3.1 and so forth until I come upto 2.3.3 version or I can apply > > directly from 2.2.1 to 2.3.3? > > > > ...Jagruti > > -----Original Message----- > > From: owner-petsc-users at mcs.anl.gov > > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Satish Balay > > Sent: Tuesday, October 09, 2007 11:29 > > To: petsc-users at mcs.anl.gov > > Subject: Re: Compiled information for Petsc 2.3.2-p10 > > > > > > It is possible that we didn't document all the changes between releases. > > > > But the recommended approach is to look at changes file for each > > version after 2.2.1 - and apply these changes. > > > > And if your code is in C - then the compiler will help in finding > > any undocumented changes in prototypes.. etc.. > > > > BTW: the current release version is 2.3.3 [so you might want to > > upgrade to this version instead of 2.3.2] > > > > Satish > > > > On Tue, 9 Oct 2007, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > > > > > > > > > > Hello there, > > > > > > I am trying to port my application from petsc 2.2.1 to petsc > > 2.3.2-p10. > > > Has anybody compiled > > > The real life changes needed to make other than listed in the > > > changes file? > > > > > > I really appreciate if you have hands on do's and don'ts for > > > porting from petsc-2.2.1 to Petsc 2.3.2-p10. > > > > > > Looking forward to positive replies. > > > > > > Thanks and Best Regards, > > > > > > ....Jagruti > > > > > > NAWC > > > > > > > > > > -- > Lisandro Dalc?n > --------------- > Centro Internacional de M?todos Computacionales en Ingenier?a (CIMEC) > Instituto de Desarrollo Tecnol?gico para la Industria Qu?mica (INTEC) > Consejo Nacional de Investigaciones Cient?ficas y T?cnicas (CONICET) > PTLC - G?emes 3450, (3000) Santa Fe, Argentina > Tel/Fax: +54-(0)342-451.1594 > > From bsmith at mcs.anl.gov Tue Oct 9 19:41:22 2007 From: bsmith at mcs.anl.gov (Barry Smith) Date: Tue, 9 Oct 2007 19:41:22 -0500 (CDT) Subject: Compiled information for Petsc 2.3.2-p10 In-Reply-To: References: Message-ID: Sounds like you are NOT using the handy PETSc makefiles; in the PETSc main directory you can do make showlinklibs and it will show you what libraries you need to list in your makefile Barry On Tue, 9 Oct 2007, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > Barry, > > When I compile my application it fails with a linking "ld" error that it can not > Find -lmpich. With this new install, I have mpich2-1.0.5p4. Where do I change from mpich to mpich2 linking? > > Thanks, > > ...Jagruti > > -----Original Message----- > From: owner-petsc-users at mcs.anl.gov [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Barry Smith > Sent: Tuesday, October 09, 2007 16:47 > To: petsc-users at mcs.anl.gov > Subject: RE: Compiled information for Petsc 2.3.2-p10 > > > Strange, never seen that before. It is likely you can just ignore it. > > Barry > > > > > On Tue, 9 Oct 2007, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > > > > Hello Licendro, > > > > Here is the first question after the installation of petsc-2.3.3-p6. > > > > I ran: > > Config/config.py --with-cc=gcc --with-fc=g77 > > --download-f-blas-lapack=1 --download-mpich=1 > > > > /usr/local/bin/make all test /* run gnu make > > > > I get the following status error: > > > > Running test examples to verify correct installation c/C++ example > > src/snes/examples/tutorials/ex19 run successfully with 1 MPI process > > > > c/C++ example src/snes/examples/tutorials/ex19 run successfully with 2 > > MPI process > > > > Graphics example src/snes/examples/tutorials/ex19 run successfully > > with 1 MPI process > > > > /bin/sh: test: argument expected > > make[3]: [ex19.rm] Error 1 (ignored) > > /bin/sh: test: argument expected > > make[3]: [ex19.rm] Error 1 (ignored) > > /bin/sh: test: argument expected > > make[3]: [ex19.rm] Error 1 (ignored) > > /bin/sh: test: argument expected > > make[3]: [ex19.rm] Error 1 (ignored) > > Fortran example src/snes/examples/tutorials/ex5f run successfully with > > 1 MPI process > > /bin/sh: test: argument expected > > make[3]: [ex5f.rm] Error 1 (ignored) > > /bin/sh: test: argument expected > > make[3]: [ex5f.rm] Error 1 (ignored) > > /bin/sh: test: argument expected > > make[3]: [ex5f.rm] Error 1 (ignored) > > /bin/sh: test: argument expected > > make[3]: [ex5f.rm] Error 1 (ignored) > > Completed test examples > > > > Is this the correct installation? Why am I getting this error messages? > > > > Thanks, > > > > ....Jagruti > > > > -----Original Message----- > > From: owner-petsc-users at mcs.anl.gov > > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Lisandro Dalcin > > Sent: Tuesday, October 09, 2007 12:31 > > To: petsc-users at mcs.anl.gov > > Subject: Re: Compiled information for Petsc 2.3.2-p10 > > > > I want to suggest you to update to 2.3.3. Of course, you have the extra trouble of figuring out yourself the cummulative changes since 2.2.1. > > > > Go ahead updating to 2.3.3, and I can help you at any point you have trouble. > > > > > > > > On 10/9/07, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > > > > > Do you mean that I should apply changes from 2.2.1 to 2.3.0 and then > > > 2.3.1 and so forth until I come upto 2.3.3 version or I can apply > > > directly from 2.2.1 to 2.3.3? > > > > > > ...Jagruti > > > -----Original Message----- > > > From: owner-petsc-users at mcs.anl.gov > > > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Satish Balay > > > Sent: Tuesday, October 09, 2007 11:29 > > > To: petsc-users at mcs.anl.gov > > > Subject: Re: Compiled information for Petsc 2.3.2-p10 > > > > > > > > > It is possible that we didn't document all the changes between releases. > > > > > > But the recommended approach is to look at changes file for each > > > version after 2.2.1 - and apply these changes. > > > > > > And if your code is in C - then the compiler will help in finding > > > any undocumented changes in prototypes.. etc.. > > > > > > BTW: the current release version is 2.3.3 [so you might want to > > > upgrade to this version instead of 2.3.2] > > > > > > Satish > > > > > > On Tue, 9 Oct 2007, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > > > > > > > > > > > > > > Hello there, > > > > > > > > I am trying to port my application from petsc 2.2.1 to petsc > > > 2.3.2-p10. > > > > Has anybody compiled > > > > The real life changes needed to make other than listed in the > > > > changes file? > > > > > > > > I really appreciate if you have hands on do's and don'ts for > > > > porting from petsc-2.2.1 to Petsc 2.3.2-p10. > > > > > > > > Looking forward to positive replies. > > > > > > > > Thanks and Best Regards, > > > > > > > > ....Jagruti > > > > > > > > NAWC > > > > > > > > > > > > > > > > -- > > Lisandro Dalc?n > > --------------- > > Centro Internacional de M?todos Computacionales en Ingenier?a (CIMEC) > > Instituto de Desarrollo Tecnol?gico para la Industria Qu?mica (INTEC) > > Consejo Nacional de Investigaciones Cient?ficas y T?cnicas (CONICET) > > PTLC - G?emes 3450, (3000) Santa Fe, Argentina > > Tel/Fax: +54-(0)342-451.1594 > > > > > > From renzhengyong at gmail.com Tue Oct 9 22:23:27 2007 From: renzhengyong at gmail.com (RenZhengYong) Date: Tue, 9 Oct 2007 23:23:27 -0400 Subject: which solving methods is more efficient to my problem. In-Reply-To: References: <7ff0ee010710062327i583253a8j3fd1bc6b95f62424@mail.gmail.com> Message-ID: <4913f8f50710092023g28e8c4ekae6d13f014e17444@mail.gmail.com> Sorry for that. Actually, the simple linear elliptic (second order) PDE is adopted in my codes. And, my results tell me that the MG method is more preferred than CG-like solvers. On 10/8/07, Lisandro Dalcin wrote: > > You forgot to mention what kind of problem your PDE is, appart from > being linear. I mean, can you describe which kind of problem are you > trying to solve? > > On 10/7/07, Yujie wrote: > > Hi, everyone > > > > I am using adaptive mesh refinement to solve a group of coupled > steady-state > > linear PDEs. I will use PETSC as the solver. Did anyone have experience > > in such problem. I want to know which > > solving method is more efficient, including > > iterative method and preconditioner. > > Thanks a lot. > > > > Regards, > > Yujie > > > > > -- > Lisandro Dalc?n > --------------- > Centro Internacional de M?todos Computacionales en Ingenier?a (CIMEC) > Instituto de Desarrollo Tecnol?gico para la Industria Qu?mica (INTEC) > Consejo Nacional de Investigaciones Cient?ficas y T?cnicas (CONICET) > PTLC - G?emes 3450, (3000) Santa Fe, Argentina > Tel/Fax: +54-(0)342-451.1594 > > -- Zhengyong Ren School of Info-physics and Geomatics Engineering DiXue Building, 3-322#, Central South University, Hunan Province,China. Postal code: 410083 Email: renzhengyong at gmail.com Tel: +086013787102372. -------------- next part -------------- An HTML attachment was scrubbed... URL: From jagruti.trivedi at navy.mil Wed Oct 10 13:20:05 2007 From: jagruti.trivedi at navy.mil (Trivedi, Jagruti CIV 470000D, 474300D) Date: Wed, 10 Oct 2007 11:20:05 -0700 Subject: Symbol SNESGetNumberLinearIterations not defined In-Reply-To: References: Message-ID: Hello There, In compiling my application code ( petsc-2.3.3-p6), I get the following symbol SNESGetNumberLinearIterations, first defined. I checked the 2.3.3 changes file and It does mention of adding this function. I checked the manualpages and I could not find the function under SNES directory. Am I suppose to include any other header file? I have included "petscsnes.h" file in my code. Best Regards, ...Jagruti -----Original Message----- From: owner-petsc-users at mcs.anl.gov [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Lisandro Dalcin Sent: Tuesday, October 09, 2007 17:08 To: petsc-users at mcs.anl.gov Subject: Re: Compiled information for Petsc 2.3.2-p10 Perhaps PETSc is not internally calling '/usr/local/bin/make' ?? Was '/usr/local/bin/test' in front of your $PATH when you configured PETSc? As Barry said, you can just ignore this, unless you have problems building your own source. In that case, please send 'configure.log' file. On 10/9/07, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > Hello Licendro, > > Here is the first question after the installation of petsc-2.3.3-p6. > > I ran: > Config/config.py --with-cc=gcc --with-fc=g77 > --download-f-blas-lapack=1 --download-mpich=1 > > /usr/local/bin/make all test /* run gnu make > > I get the following status error: > > Running test examples to verify correct installation c/C++ example > src/snes/examples/tutorials/ex19 run successfully with 1 MPI process > > c/C++ example src/snes/examples/tutorials/ex19 run successfully with 2 > MPI process > > Graphics example src/snes/examples/tutorials/ex19 run successfully > with 1 MPI process > > /bin/sh: test: argument expected > make[3]: [ex19.rm] Error 1 (ignored) > /bin/sh: test: argument expected > make[3]: [ex19.rm] Error 1 (ignored) > /bin/sh: test: argument expected > make[3]: [ex19.rm] Error 1 (ignored) > /bin/sh: test: argument expected > make[3]: [ex19.rm] Error 1 (ignored) > Fortran example src/snes/examples/tutorials/ex5f run successfully with > 1 MPI process > /bin/sh: test: argument expected > make[3]: [ex5f.rm] Error 1 (ignored) > /bin/sh: test: argument expected > make[3]: [ex5f.rm] Error 1 (ignored) > /bin/sh: test: argument expected > make[3]: [ex5f.rm] Error 1 (ignored) > /bin/sh: test: argument expected > make[3]: [ex5f.rm] Error 1 (ignored) > Completed test examples > > Is this the correct installation? Why am I getting this error messages? > > Thanks, > > ....Jagruti > > -----Original Message----- > From: owner-petsc-users at mcs.anl.gov > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Lisandro Dalcin > Sent: Tuesday, October 09, 2007 12:31 > To: petsc-users at mcs.anl.gov > Subject: Re: Compiled information for Petsc 2.3.2-p10 > > I want to suggest you to update to 2.3.3. Of course, you have the extra trouble of figuring out yourself the cummulative changes since 2.2.1. > > Go ahead updating to 2.3.3, and I can help you at any point you have trouble. > > > > On 10/9/07, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > > > Do you mean that I should apply changes from 2.2.1 to 2.3.0 and then > > 2.3.1 and so forth until I come upto 2.3.3 version or I can apply > > directly from 2.2.1 to 2.3.3? > > > > ...Jagruti > > -----Original Message----- > > From: owner-petsc-users at mcs.anl.gov > > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Satish Balay > > Sent: Tuesday, October 09, 2007 11:29 > > To: petsc-users at mcs.anl.gov > > Subject: Re: Compiled information for Petsc 2.3.2-p10 > > > > > > It is possible that we didn't document all the changes between releases. > > > > But the recommended approach is to look at changes file for each > > version after 2.2.1 - and apply these changes. > > > > And if your code is in C - then the compiler will help in finding > > any undocumented changes in prototypes.. etc.. > > > > BTW: the current release version is 2.3.3 [so you might want to > > upgrade to this version instead of 2.3.2] > > > > Satish > > > > On Tue, 9 Oct 2007, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > > > > > > > > > > Hello there, > > > > > > I am trying to port my application from petsc 2.2.1 to petsc > > 2.3.2-p10. > > > Has anybody compiled > > > The real life changes needed to make other than listed in the > > > changes file? > > > > > > I really appreciate if you have hands on do's and don'ts for > > > porting from petsc-2.2.1 to Petsc 2.3.2-p10. > > > > > > Looking forward to positive replies. > > > > > > Thanks and Best Regards, > > > > > > ....Jagruti > > > > > > NAWC > > > > > > > > > > -- > Lisandro Dalc?n > --------------- > Centro Internacional de M?todos Computacionales en Ingenier?a (CIMEC) > Instituto de Desarrollo Tecnol?gico para la Industria Qu?mica (INTEC) > Consejo Nacional de Investigaciones Cient?ficas y T?cnicas (CONICET) > PTLC - G?emes 3450, (3000) Santa Fe, Argentina > Tel/Fax: +54-(0)342-451.1594 > > -- Lisandro Dalc?n --------------- Centro Internacional de M?todos Computacionales en Ingenier?a (CIMEC) Instituto de Desarrollo Tecnol?gico para la Industria Qu?mica (INTEC) Consejo Nacional de Investigaciones Cient?ficas y T?cnicas (CONICET) PTLC - G?emes 3450, (3000) Santa Fe, Argentina Tel/Fax: +54-(0)342-451.1594 From bsmith at mcs.anl.gov Wed Oct 10 13:30:14 2007 From: bsmith at mcs.anl.gov (Barry Smith) Date: Wed, 10 Oct 2007 13:30:14 -0500 (CDT) Subject: Symbol SNESGetNumberLinearIterations not defined In-Reply-To: References: Message-ID: SNESGetLinearSolveIterations() On Wed, 10 Oct 2007, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > > Hello There, > > In compiling my application code ( petsc-2.3.3-p6), I get the following symbol > SNESGetNumberLinearIterations, first defined. I checked the 2.3.3 changes file and > It does mention of adding this function. I checked the manualpages and I could not find the function under SNES directory. Am I suppose to include any other header file? > > I have included "petscsnes.h" file in my code. > > Best Regards, > > ...Jagruti > > -----Original Message----- > From: owner-petsc-users at mcs.anl.gov [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Lisandro Dalcin > Sent: Tuesday, October 09, 2007 17:08 > To: petsc-users at mcs.anl.gov > Subject: Re: Compiled information for Petsc 2.3.2-p10 > > Perhaps PETSc is not internally calling '/usr/local/bin/make' ?? Was '/usr/local/bin/test' in front of your $PATH when you configured PETSc? > > As Barry said, you can just ignore this, unless you have problems building your own source. In that case, please send 'configure.log' > file. > > > On 10/9/07, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > > > Hello Licendro, > > > > Here is the first question after the installation of petsc-2.3.3-p6. > > > > I ran: > > Config/config.py --with-cc=gcc --with-fc=g77 > > --download-f-blas-lapack=1 --download-mpich=1 > > > > /usr/local/bin/make all test /* run gnu make > > > > I get the following status error: > > > > Running test examples to verify correct installation c/C++ example > > src/snes/examples/tutorials/ex19 run successfully with 1 MPI process > > > > c/C++ example src/snes/examples/tutorials/ex19 run successfully with 2 > > MPI process > > > > Graphics example src/snes/examples/tutorials/ex19 run successfully > > with 1 MPI process > > > > /bin/sh: test: argument expected > > make[3]: [ex19.rm] Error 1 (ignored) > > /bin/sh: test: argument expected > > make[3]: [ex19.rm] Error 1 (ignored) > > /bin/sh: test: argument expected > > make[3]: [ex19.rm] Error 1 (ignored) > > /bin/sh: test: argument expected > > make[3]: [ex19.rm] Error 1 (ignored) > > Fortran example src/snes/examples/tutorials/ex5f run successfully with > > 1 MPI process > > /bin/sh: test: argument expected > > make[3]: [ex5f.rm] Error 1 (ignored) > > /bin/sh: test: argument expected > > make[3]: [ex5f.rm] Error 1 (ignored) > > /bin/sh: test: argument expected > > make[3]: [ex5f.rm] Error 1 (ignored) > > /bin/sh: test: argument expected > > make[3]: [ex5f.rm] Error 1 (ignored) > > Completed test examples > > > > Is this the correct installation? Why am I getting this error messages? > > > > Thanks, > > > > ....Jagruti > > > > -----Original Message----- > > From: owner-petsc-users at mcs.anl.gov > > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Lisandro Dalcin > > Sent: Tuesday, October 09, 2007 12:31 > > To: petsc-users at mcs.anl.gov > > Subject: Re: Compiled information for Petsc 2.3.2-p10 > > > > I want to suggest you to update to 2.3.3. Of course, you have the extra trouble of figuring out yourself the cummulative changes since 2.2.1. > > > > Go ahead updating to 2.3.3, and I can help you at any point you have trouble. > > > > > > > > On 10/9/07, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > > > > > Do you mean that I should apply changes from 2.2.1 to 2.3.0 and then > > > 2.3.1 and so forth until I come upto 2.3.3 version or I can apply > > > directly from 2.2.1 to 2.3.3? > > > > > > ...Jagruti > > > -----Original Message----- > > > From: owner-petsc-users at mcs.anl.gov > > > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Satish Balay > > > Sent: Tuesday, October 09, 2007 11:29 > > > To: petsc-users at mcs.anl.gov > > > Subject: Re: Compiled information for Petsc 2.3.2-p10 > > > > > > > > > It is possible that we didn't document all the changes between releases. > > > > > > But the recommended approach is to look at changes file for each > > > version after 2.2.1 - and apply these changes. > > > > > > And if your code is in C - then the compiler will help in finding > > > any undocumented changes in prototypes.. etc.. > > > > > > BTW: the current release version is 2.3.3 [so you might want to > > > upgrade to this version instead of 2.3.2] > > > > > > Satish > > > > > > On Tue, 9 Oct 2007, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > > > > > > > > > > > > > > Hello there, > > > > > > > > I am trying to port my application from petsc 2.2.1 to petsc > > > 2.3.2-p10. > > > > Has anybody compiled > > > > The real life changes needed to make other than listed in the > > > > changes file? > > > > > > > > I really appreciate if you have hands on do's and don'ts for > > > > porting from petsc-2.2.1 to Petsc 2.3.2-p10. > > > > > > > > Looking forward to positive replies. > > > > > > > > Thanks and Best Regards, > > > > > > > > ....Jagruti > > > > > > > > NAWC > > > > > > > > > > > > > > > > -- > > Lisandro Dalc?n > > --------------- > > Centro Internacional de M?todos Computacionales en Ingenier?a (CIMEC) > > Instituto de Desarrollo Tecnol?gico para la Industria Qu?mica (INTEC) > > Consejo Nacional de Investigaciones Cient?ficas y T?cnicas (CONICET) > > PTLC - G?emes 3450, (3000) Santa Fe, Argentina > > Tel/Fax: +54-(0)342-451.1594 > > > > > > > -- > Lisandro Dalc?n > --------------- > Centro Internacional de M?todos Computacionales en Ingenier?a (CIMEC) Instituto de Desarrollo Tecnol?gico para la Industria Qu?mica (INTEC) Consejo Nacional de Investigaciones Cient?ficas y T?cnicas (CONICET) PTLC - G?emes 3450, (3000) Santa Fe, Argentina > Tel/Fax: +54-(0)342-451.1594 > > From mfhoel at ifi.uio.no Wed Oct 10 13:34:48 2007 From: mfhoel at ifi.uio.no (Mads Hoel) Date: Wed, 10 Oct 2007 20:34:48 +0200 Subject: Compiled information for Petsc 2.3.2-p10 In-Reply-To: References: Message-ID: P? Wed, 10 Oct 2007 02:41:22 +0200, skrev Barry Smith : Hi all, "make showlinklibs" does not seem to be defined in the makefile(s). It has probably been renamed to "make getlinklibs", since that works. Might want to check out "make getmpilinklibs" also. They are all defined in ${PETSC_DIR}/bmake/common/rules. > > Sounds like you are NOT using the handy PETSc makefiles; in the PETSc > main directory you can do > make showlinklibs > and it will show you what libraries you need to list in your makefile > > Barry > > > On Tue, 9 Oct 2007, Trivedi, Jagruti CIV 470000D, 474300D wrote: > >> >> Barry, >> >> When I compile my application it fails with a linking "ld" error that >> it can not >> Find -lmpich. With this new install, I have mpich2-1.0.5p4. Where do I >> change from mpich to mpich2 linking? >> >> Thanks, >> >> ...Jagruti >> >> -----Original Message----- >> From: owner-petsc-users at mcs.anl.gov >> [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Barry Smith >> Sent: Tuesday, October 09, 2007 16:47 >> To: petsc-users at mcs.anl.gov >> Subject: RE: Compiled information for Petsc 2.3.2-p10 >> >> >> Strange, never seen that before. It is likely you can just ignore it. >> >> Barry >> >> >> >> >> On Tue, 9 Oct 2007, Trivedi, Jagruti CIV 470000D, 474300D wrote: >> >> > >> > Hello Licendro, >> > >> > Here is the first question after the installation of petsc-2.3.3-p6. >> > >> > I ran: >> > Config/config.py --with-cc=gcc --with-fc=g77 >> > --download-f-blas-lapack=1 --download-mpich=1 >> > >> > /usr/local/bin/make all test /* run gnu make >> > >> > I get the following status error: >> > >> > Running test examples to verify correct installation c/C++ example >> > src/snes/examples/tutorials/ex19 run successfully with 1 MPI process >> > >> > c/C++ example src/snes/examples/tutorials/ex19 run successfully with 2 >> > MPI process >> > >> > Graphics example src/snes/examples/tutorials/ex19 run successfully >> > with 1 MPI process >> > >> > /bin/sh: test: argument expected >> > make[3]: [ex19.rm] Error 1 (ignored) >> > /bin/sh: test: argument expected >> > make[3]: [ex19.rm] Error 1 (ignored) >> > /bin/sh: test: argument expected >> > make[3]: [ex19.rm] Error 1 (ignored) >> > /bin/sh: test: argument expected >> > make[3]: [ex19.rm] Error 1 (ignored) >> > Fortran example src/snes/examples/tutorials/ex5f run successfully with >> > 1 MPI process >> > /bin/sh: test: argument expected >> > make[3]: [ex5f.rm] Error 1 (ignored) >> > /bin/sh: test: argument expected >> > make[3]: [ex5f.rm] Error 1 (ignored) >> > /bin/sh: test: argument expected >> > make[3]: [ex5f.rm] Error 1 (ignored) >> > /bin/sh: test: argument expected >> > make[3]: [ex5f.rm] Error 1 (ignored) >> > Completed test examples >> > >> > Is this the correct installation? Why am I getting this error >> messages? >> > >> > Thanks, >> > >> > ....Jagruti >> > >> > -----Original Message----- >> > From: owner-petsc-users at mcs.anl.gov >> > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Lisandro Dalcin >> > Sent: Tuesday, October 09, 2007 12:31 >> > To: petsc-users at mcs.anl.gov >> > Subject: Re: Compiled information for Petsc 2.3.2-p10 >> > >> > I want to suggest you to update to 2.3.3. Of course, you have the >> extra trouble of figuring out yourself the cummulative changes since >> 2.2.1. >> > >> > Go ahead updating to 2.3.3, and I can help you at any point you have >> trouble. >> > >> > >> > >> > On 10/9/07, Trivedi, Jagruti CIV 470000D, 474300D >> wrote: >> > > >> > > Do you mean that I should apply changes from 2.2.1 to 2.3.0 and then >> > > 2.3.1 and so forth until I come upto 2.3.3 version or I can apply >> > > directly from 2.2.1 to 2.3.3? >> > > >> > > ...Jagruti >> > > -----Original Message----- >> > > From: owner-petsc-users at mcs.anl.gov >> > > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Satish Balay >> > > Sent: Tuesday, October 09, 2007 11:29 >> > > To: petsc-users at mcs.anl.gov >> > > Subject: Re: Compiled information for Petsc 2.3.2-p10 >> > > >> > > >> > > It is possible that we didn't document all the changes between >> releases. >> > > >> > > But the recommended approach is to look at changes file for each >> > > version after 2.2.1 - and apply these changes. >> > > >> > > And if your code is in C - then the compiler will help in finding >> > > any undocumented changes in prototypes.. etc.. >> > > >> > > BTW: the current release version is 2.3.3 [so you might want to >> > > upgrade to this version instead of 2.3.2] >> > > >> > > Satish >> > > >> > > On Tue, 9 Oct 2007, Trivedi, Jagruti CIV 470000D, 474300D wrote: >> > > >> > > > >> > > > >> > > > Hello there, >> > > > >> > > > I am trying to port my application from petsc 2.2.1 to petsc >> > > 2.3.2-p10. >> > > > Has anybody compiled >> > > > The real life changes needed to make other than listed in the >> > > > changes file? >> > > > >> > > > I really appreciate if you have hands on do's and don'ts for >> > > > porting from petsc-2.2.1 to Petsc 2.3.2-p10. >> > > > >> > > > Looking forward to positive replies. >> > > > >> > > > Thanks and Best Regards, >> > > > >> > > > ....Jagruti >> > > > >> > > > NAWC >> > > > >> > > >> > > >> > >> > >> > -- >> > Lisandro Dalc?n >> > --------------- >> > Centro Internacional de M?todos Computacionales en Ingenier?a (CIMEC) >> > Instituto de Desarrollo Tecnol?gico para la Industria Qu?mica (INTEC) >> > Consejo Nacional de Investigaciones Cient?ficas y T?cnicas (CONICET) >> > PTLC - G?emes 3450, (3000) Santa Fe, Argentina >> > Tel/Fax: +54-(0)342-451.1594 >> > >> > >> -- Mads Hoel From jagruti.trivedi at navy.mil Wed Oct 10 13:44:35 2007 From: jagruti.trivedi at navy.mil (Trivedi, Jagruti CIV 470000D, 474300D) Date: Wed, 10 Oct 2007 11:44:35 -0700 Subject: Symbol SNESGetNumberLinearIterations not defined In-Reply-To: References: Message-ID: Hi Barry, Thank you. Do you want me to replace this new function for "SNESGetNumberLinearIterations"?. Can you explain why the Changes says, this function SNESGetNumberLinearIterations() function is added? I do not want to change the interface if possible. Best Reagrds, ...Jagruti -----Original Message----- From: owner-petsc-users at mcs.anl.gov [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Barry Smith Sent: Wednesday, October 10, 2007 11:30 To: petsc-users at mcs.anl.gov Subject: Re: Symbol SNESGetNumberLinearIterations not defined SNESGetLinearSolveIterations() On Wed, 10 Oct 2007, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > > Hello There, > > In compiling my application code ( petsc-2.3.3-p6), I get the > following symbol SNESGetNumberLinearIterations, first defined. I > checked the 2.3.3 changes file and It does mention of adding this function. I checked the manualpages and I could not find the function under SNES directory. Am I suppose to include any other header file? > > I have included "petscsnes.h" file in my code. > > Best Regards, > > ...Jagruti > > -----Original Message----- > From: owner-petsc-users at mcs.anl.gov > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Lisandro Dalcin > Sent: Tuesday, October 09, 2007 17:08 > To: petsc-users at mcs.anl.gov > Subject: Re: Compiled information for Petsc 2.3.2-p10 > > Perhaps PETSc is not internally calling '/usr/local/bin/make' ?? Was '/usr/local/bin/test' in front of your $PATH when you configured PETSc? > > As Barry said, you can just ignore this, unless you have problems building your own source. In that case, please send 'configure.log' > file. > > > On 10/9/07, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > > > Hello Licendro, > > > > Here is the first question after the installation of petsc-2.3.3-p6. > > > > I ran: > > Config/config.py --with-cc=gcc --with-fc=g77 > > --download-f-blas-lapack=1 --download-mpich=1 > > > > /usr/local/bin/make all test /* run gnu make > > > > I get the following status error: > > > > Running test examples to verify correct installation c/C++ example > > src/snes/examples/tutorials/ex19 run successfully with 1 MPI process > > > > c/C++ example src/snes/examples/tutorials/ex19 run successfully with > > 2 MPI process > > > > Graphics example src/snes/examples/tutorials/ex19 run successfully > > with 1 MPI process > > > > /bin/sh: test: argument expected > > make[3]: [ex19.rm] Error 1 (ignored) > > /bin/sh: test: argument expected > > make[3]: [ex19.rm] Error 1 (ignored) > > /bin/sh: test: argument expected > > make[3]: [ex19.rm] Error 1 (ignored) > > /bin/sh: test: argument expected > > make[3]: [ex19.rm] Error 1 (ignored) Fortran example > > src/snes/examples/tutorials/ex5f run successfully with > > 1 MPI process > > /bin/sh: test: argument expected > > make[3]: [ex5f.rm] Error 1 (ignored) > > /bin/sh: test: argument expected > > make[3]: [ex5f.rm] Error 1 (ignored) > > /bin/sh: test: argument expected > > make[3]: [ex5f.rm] Error 1 (ignored) > > /bin/sh: test: argument expected > > make[3]: [ex5f.rm] Error 1 (ignored) Completed test examples > > > > Is this the correct installation? Why am I getting this error messages? > > > > Thanks, > > > > ....Jagruti > > > > -----Original Message----- > > From: owner-petsc-users at mcs.anl.gov > > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Lisandro Dalcin > > Sent: Tuesday, October 09, 2007 12:31 > > To: petsc-users at mcs.anl.gov > > Subject: Re: Compiled information for Petsc 2.3.2-p10 > > > > I want to suggest you to update to 2.3.3. Of course, you have the extra trouble of figuring out yourself the cummulative changes since 2.2.1. > > > > Go ahead updating to 2.3.3, and I can help you at any point you have trouble. > > > > > > > > On 10/9/07, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > > > > > Do you mean that I should apply changes from 2.2.1 to 2.3.0 and > > > then > > > 2.3.1 and so forth until I come upto 2.3.3 version or I can apply > > > directly from 2.2.1 to 2.3.3? > > > > > > ...Jagruti > > > -----Original Message----- > > > From: owner-petsc-users at mcs.anl.gov > > > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Satish Balay > > > Sent: Tuesday, October 09, 2007 11:29 > > > To: petsc-users at mcs.anl.gov > > > Subject: Re: Compiled information for Petsc 2.3.2-p10 > > > > > > > > > It is possible that we didn't document all the changes between releases. > > > > > > But the recommended approach is to look at changes file for each > > > version after 2.2.1 - and apply these changes. > > > > > > And if your code is in C - then the compiler will help in finding > > > any undocumented changes in prototypes.. etc.. > > > > > > BTW: the current release version is 2.3.3 [so you might want to > > > upgrade to this version instead of 2.3.2] > > > > > > Satish > > > > > > On Tue, 9 Oct 2007, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > > > > > > > > > > > > > > Hello there, > > > > > > > > I am trying to port my application from petsc 2.2.1 to petsc > > > 2.3.2-p10. > > > > Has anybody compiled > > > > The real life changes needed to make other than listed in the > > > > changes file? > > > > > > > > I really appreciate if you have hands on do's and don'ts for > > > > porting from petsc-2.2.1 to Petsc 2.3.2-p10. > > > > > > > > Looking forward to positive replies. > > > > > > > > Thanks and Best Regards, > > > > > > > > ....Jagruti > > > > > > > > NAWC > > > > > > > > > > > > > > > > -- > > Lisandro Dalc?n > > --------------- > > Centro Internacional de M?todos Computacionales en Ingenier?a > > (CIMEC) Instituto de Desarrollo Tecnol?gico para la Industria > > Qu?mica (INTEC) Consejo Nacional de Investigaciones Cient?ficas y > > T?cnicas (CONICET) PTLC - G?emes 3450, (3000) Santa Fe, Argentina > > Tel/Fax: +54-(0)342-451.1594 > > > > > > > -- > Lisandro Dalc?n > --------------- > Centro Internacional de M?todos Computacionales en Ingenier?a (CIMEC) > Instituto de Desarrollo Tecnol?gico para la Industria Qu?mica (INTEC) > Consejo Nacional de Investigaciones Cient?ficas y T?cnicas (CONICET) > PTLC - G?emes 3450, (3000) Santa Fe, Argentina > Tel/Fax: +54-(0)342-451.1594 > > From balay at mcs.anl.gov Wed Oct 10 13:54:41 2007 From: balay at mcs.anl.gov (Satish Balay) Date: Wed, 10 Oct 2007 13:54:41 -0500 (CDT) Subject: Symbol SNESGetNumberLinearIterations not defined In-Reply-To: References: Message-ID: On Wed, 10 Oct 2007, Trivedi, Jagruti CIV 470000D, 474300D wrote: > Thank you. Do you want me to replace this new function for > "SNESGetNumberLinearIterations"?. yes. > Can you explain why the Changes says, this function > SNESGetNumberLinearIterations() function is added? I do not want to > change the interface if possible. The changes file you are refering to is petsc-2.0.7. so SNESGetNumberLinearIterations() was added in petsc-2.0.7. This function is changed to SNESGetLinearSolveIterations() [perhaps in 2.3.3 - but didn't get documented in the changes file for that version] Satish From jagruti.trivedi at navy.mil Wed Oct 10 15:47:09 2007 From: jagruti.trivedi at navy.mil (Trivedi, Jagruti CIV 470000D, 474300D) Date: Wed, 10 Oct 2007 13:47:09 -0700 Subject: Symbol SNESGetNumberLinearIterations not defined In-Reply-To: References: Message-ID: Hello Everybody, Thanks to every one who helped. I successfully ported ( compiled and linked without warning ) my application on PetSc 2.3.3-p6. I have yet to test it so that it works as it suppose to also. Best Regards, ....Jagruti -----Original Message----- From: owner-petsc-users at mcs.anl.gov [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Satish Balay Sent: Wednesday, October 10, 2007 11:55 To: petsc-users at mcs.anl.gov Subject: RE: Symbol SNESGetNumberLinearIterations not defined On Wed, 10 Oct 2007, Trivedi, Jagruti CIV 470000D, 474300D wrote: > Thank you. Do you want me to replace this new function for > "SNESGetNumberLinearIterations"?. yes. > Can you explain why the Changes says, this function > SNESGetNumberLinearIterations() function is added? I do not want to > change the interface if possible. The changes file you are refering to is petsc-2.0.7. so SNESGetNumberLinearIterations() was added in petsc-2.0.7. This function is changed to SNESGetLinearSolveIterations() [perhaps in 2.3.3 - but didn't get documented in the changes file for that version] Satish From mfhoel at ifi.uio.no Thu Oct 11 07:37:38 2007 From: mfhoel at ifi.uio.no (Mads Hoel) Date: Thu, 11 Oct 2007 14:37:38 +0200 Subject: ISLocalToGlobalMappingApply(mapping,m,input,output); and alike "functions" Message-ID: Hi all, I been having trouble with this function today. I called it from a function returning a void. Later found out it was actually a macro with a "return PetscError(__LINE__,__FUNCT__,__FILE__,__SDIR__,n,1,s,a1,a2,a3);" statement in it. Which makes it not possible to call ISLocalToGlobalMappingApply from a function returning a void, when compiling with g++ compiler. Or perhaps it may lead to difficult to find error in functions returning another type than an int, for example an object with a constructor accepting an int. Thanks -- Mads Hoel From keita at cray.com Thu Oct 11 16:46:33 2007 From: keita at cray.com (Keita Teranishi) Date: Thu, 11 Oct 2007 16:46:33 -0500 Subject: Preconditioner for bicg Message-ID: <925346A443D4E340BEB20248BAFCDBDF02A11CEC@CFEVS1-IP.americas.cray.com> Hi, I'd like to make sure preconditioners supported for bicg. I found bicg only accepts preconditioners that is interfaced through PCApplyTranspose. For this reason, I see several preconditioners fails with bicg. Thanks, ================================ Keita Teranishi Math Software Group Cray Inc. keita at cray.com ================================ -------------- next part -------------- An HTML attachment was scrubbed... URL: From bsmith at mcs.anl.gov Thu Oct 11 16:50:15 2007 From: bsmith at mcs.anl.gov (Barry Smith) Date: Thu, 11 Oct 2007 16:50:15 -0500 (CDT) Subject: Preconditioner for bicg In-Reply-To: <925346A443D4E340BEB20248BAFCDBDF02A11CEC@CFEVS1-IP.americas.cray.com> References: <925346A443D4E340BEB20248BAFCDBDF02A11CEC@CFEVS1-IP.americas.cray.com> Message-ID: For some preconditioners it is simply not easily possible to apply the transpose of the preconditioner. Even multigrid requires some work to do this. Barry On Thu, 11 Oct 2007, Keita Teranishi wrote: > Hi, > > > > I'd like to make sure preconditioners supported for bicg. I found bicg only accepts preconditioners that is interfaced through PCApplyTranspose. > > For this reason, I see several preconditioners fails with bicg. > > > > Thanks, > > ================================ > Keita Teranishi > Math Software Group > Cray Inc. > keita at cray.com > ================================ > > > > From keita at cray.com Thu Oct 11 17:07:53 2007 From: keita at cray.com (Keita Teranishi) Date: Thu, 11 Oct 2007 17:07:53 -0500 Subject: Preconditioner for bicg In-Reply-To: References: <925346A443D4E340BEB20248BAFCDBDF02A11CEC@CFEVS1-IP.americas.cray.com> Message-ID: <925346A443D4E340BEB20248BAFCDBDF02A11D23@CFEVS1-IP.americas.cray.com> Hi Barry, Could you suggest any other KSP solvers that have the same type of problems? Thanks, ================================ Keita Teranishi Math Software Group Cray Inc. keita at cray.com ================================ -----Original Message----- From: owner-petsc-users at mcs.anl.gov [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Barry Smith Sent: Thursday, October 11, 2007 4:50 PM To: petsc-users at mcs.anl.gov Subject: Re: Preconditioner for bicg For some preconditioners it is simply not easily possible to apply the transpose of the preconditioner. Even multigrid requires some work to do this. Barry On Thu, 11 Oct 2007, Keita Teranishi wrote: > Hi, > > > > I'd like to make sure preconditioners supported for bicg. I found bicg only accepts preconditioners that is interfaced through PCApplyTranspose. > > For this reason, I see several preconditioners fails with bicg. > > > > Thanks, > > ================================ > Keita Teranishi > Math Software Group > Cray Inc. > keita at cray.com > ================================ > > > > From jagruti.trivedi at navy.mil Thu Oct 11 18:35:01 2007 From: jagruti.trivedi at navy.mil (Trivedi, Jagruti CIV 470000D, 474300D) Date: Thu, 11 Oct 2007 16:35:01 -0700 Subject: Version of Mpi and fblaslapack Message-ID: Hi there, How do you find out which version of mpi and fblaslapack you need with which version of PetSc? I am interested in PetSc 2.2.1. I need to install PetSc 2.2.1 to recreate the working scenerio for my Application. My application compiled and linked ok with PetSc 2.3.3 but it crashes so I need to duplicate my Application working on PetSc 2.2.1. If I run config/configure.py with download option, it downloads the latest version of mpi and fblaslapak. B/R, ...Jagruti -------------- next part -------------- An HTML attachment was scrubbed... URL: From dalcinl at gmail.com Thu Oct 11 18:54:48 2007 From: dalcinl at gmail.com (Lisandro Dalcin) Date: Thu, 11 Oct 2007 20:54:48 -0300 Subject: Version of Mpi and fblaslapack In-Reply-To: References: Message-ID: I would suggest you to manually install MPI (MPICH2 perhpas works, if not, use MPICH1), and install blas and lapack (if you are on linux, it's far easier to look for those packages on your distro repo). But perhaps the best thing is to figure out the crash... Can you send de ouput of the crash, and perhaps run your application using valgrind? On 10/11/07, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > > > > Hi there, > > How do you find out which version of mpi and fblaslapack you need with which > version of PetSc? > > I am interested in PetSc 2.2.1. I need to install PetSc 2.2.1 to recreate > the working scenerio for my > Application. > > My application compiled and linked ok with PetSc 2.3.3 but it crashes so I > need to duplicate my > Application working on PetSc 2.2.1. > > If I run config/configure.py with download option, it downloads the latest > version of mpi and fblaslapak. > > B/R, > > ?Jagruti -- Lisandro Dalc?n --------------- Centro Internacional de M?todos Computacionales en Ingenier?a (CIMEC) Instituto de Desarrollo Tecnol?gico para la Industria Qu?mica (INTEC) Consejo Nacional de Investigaciones Cient?ficas y T?cnicas (CONICET) PTLC - G?emes 3450, (3000) Santa Fe, Argentina Tel/Fax: +54-(0)342-451.1594 From bsmith at mcs.anl.gov Thu Oct 11 19:46:50 2007 From: bsmith at mcs.anl.gov (Barry Smith) Date: Thu, 11 Oct 2007 19:46:50 -0500 (CDT) Subject: Preconditioner for bicg In-Reply-To: <925346A443D4E340BEB20248BAFCDBDF02A11D23@CFEVS1-IP.americas.cray.com> References: <925346A443D4E340BEB20248BAFCDBDF02A11CEC@CFEVS1-IP.americas.cray.com> <925346A443D4E340BEB20248BAFCDBDF02A11D23@CFEVS1-IP.americas.cray.com> Message-ID: [bsmith:~] bsmith% cd Src/petsc-dev/src/ksp/ksp/impls [bsmith:ksp/ksp/impls] bsmith% grep ApplyTranspose */*.c bicg/bicg.c: ierr = KSP_PCApplyTranspose(ksp,Rl,Zl);CHKERRQ(ierr); bicg/bicg.c: ierr = KSP_PCApplyTranspose(ksp,Rl,Zl);CHKERRQ(ierr); bicg/bicg.c: ierr = KSP_PCApplyTranspose(ksp,Rl,Zl);CHKERRQ(ierr); [bsmith:ksp/ksp/impls] bsmith% grep ApplyTranspose */*/*.c cg/cgne/cgne.c: ierr = PCApplyTransposeExists(ksp->pc,&transpose_pc);CHKERRQ(ierr); cg/cgne/cgne.c: ierr = KSP_PCApplyTranspose(ksp,T,Z);CHKERRQ(ierr); cg/cgne/cgne.c: ierr = KSP_PCApplyTranspose(ksp,T,Z);CHKERRQ(ierr); These are the two methods that require the transpose operation. I have documented both of their manual pages to indicate this. Barry On Thu, 11 Oct 2007, Keita Teranishi wrote: > Hi Barry, > > Could you suggest any other KSP solvers that have the same type of problems? > > Thanks, > ================================ > Keita Teranishi > Math Software Group > Cray Inc. > keita at cray.com > ================================ > > -----Original Message----- > From: owner-petsc-users at mcs.anl.gov [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Barry Smith > Sent: Thursday, October 11, 2007 4:50 PM > To: petsc-users at mcs.anl.gov > Subject: Re: Preconditioner for bicg > > > For some preconditioners it is simply not easily possible to apply the > transpose of the preconditioner. Even multigrid requires some work to do this. > > Barry > > > On Thu, 11 Oct 2007, Keita Teranishi wrote: > > > Hi, > > > > > > > > I'd like to make sure preconditioners supported for bicg. I found bicg only accepts preconditioners that is interfaced through PCApplyTranspose. > > > > For this reason, I see several preconditioners fails with bicg. > > > > > > > > Thanks, > > > > ================================ > > Keita Teranishi > > Math Software Group > > Cray Inc. > > keita at cray.com > > ================================ > > > > > > > > > > From balay at mcs.anl.gov Thu Oct 11 22:44:00 2007 From: balay at mcs.anl.gov (Satish Balay) Date: Thu, 11 Oct 2007 22:44:00 -0500 (CDT) Subject: Version of Mpi and fblaslapack In-Reply-To: References: Message-ID: For clarification - blas/lapack & MPI have standard interfaces - so any version of these packages should work fine with all versions of PETSc. Satish On Thu, 11 Oct 2007, Lisandro Dalcin wrote: > I would suggest you to manually install MPI (MPICH2 perhpas works, if > not, use MPICH1), and install blas and lapack (if you are on linux, > it's far easier to look for those packages on your distro repo). > > But perhaps the best thing is to figure out the crash... Can you send > de ouput of the crash, and perhaps run your application using > valgrind? > > > On 10/11/07, Trivedi, Jagruti CIV 470000D, 474300D > wrote: > > > > > > > > > > Hi there, > > > > How do you find out which version of mpi and fblaslapack you need with which > > version of PetSc? > > > > I am interested in PetSc 2.2.1. I need to install PetSc 2.2.1 to recreate > > the working scenerio for my > > Application. > > > > My application compiled and linked ok with PetSc 2.3.3 but it crashes so I > > need to duplicate my > > Application working on PetSc 2.2.1. > > > > If I run config/configure.py with download option, it downloads the latest > > version of mpi and fblaslapak. > > > > B/R, > > > > ?Jagruti > > > From mfhoel at ifi.uio.no Fri Oct 12 00:59:15 2007 From: mfhoel at ifi.uio.no (Mads Hoel) Date: Fri, 12 Oct 2007 07:59:15 +0200 Subject: Version of Mpi and fblaslapack In-Reply-To: References: Message-ID: P? Fri, 12 Oct 2007 01:35:01 +0200, skrev Trivedi, Jagruti CIV 470000D, 474300D : I found out that linking PETSc with another library (Diffpack) compiled with different versions of MPI will crash when calling InitializePetsc(...). My solution was to drop running config/configure with --download-mpi option, but have MPI point to the correct installation of MPI using the --with-mpi-dir=/path/to/mpi option. > > Hi there, > > How do you find out which version of mpi and fblaslapack you need with > which version of PetSc? > > I am interested in PetSc 2.2.1. I need to install PetSc 2.2.1 to > recreate the working scenerio for my > Application. > > My application compiled and linked ok with PetSc 2.3.3 but it crashes so > I need to duplicate my > Application working on PetSc 2.2.1. > > If I run config/configure.py with download option, it downloads the > latest version of mpi and fblaslapak. > > B/R, > > ...Jagruti -- Mads Hoel From bknaepen at ulb.ac.be Fri Oct 12 02:28:20 2007 From: bknaepen at ulb.ac.be (Knaepen Bernard) Date: Fri, 12 Oct 2007 09:28:20 +0200 (CEST) Subject: configure stalling on MacOSX Message-ID: <20071012072820.1D5D16AE@bonito.ulb.ac.be> Hello, is it normal that the ./config/configure.py stalls for a long time when building petsc-2.3.3-p6 on MacOSX? The step at which it stays stuck is: TESTING: checkSharedLinkerPaths from config.setCompilers(python/BuildSystem/config/setCompilers.py:1107) It remains there for around five minutes then continues the configure process without problem. Except that it leaves a fortran process alive that I have to manually kill later on. This process looks like: 17623 fortcom 60.8% 5:06.44 1 13 60 364K 12.1M 2.23M 57.8M 17614 ifortia32 0.0% 0:00.02 1 14 19 392K 1.96M 1.86M 29.0M 17613 ifortbin 0.0% 0:00.01 1 13 16 148K 992K 876K 28.6M when I use the top command. I am configuring with the following options: /config/configure.py --with-cc=mpicc --with-fc=mpif90 --with-hypre=1 --download-hypre=1 --with-debugging=0 COPTFLAGS=-O3 FOPTFLAGS=-O3 --with-cxx=mpicxx --with-x=0 --with-sudo=sudo The system is MacOSX 10.4.9 (latest) running on a MacPro (or Macbook Pro) and the fortran compiler is ifort 10.0.17. Regards, Bernard. From knepley at gmail.com Fri Oct 12 02:59:01 2007 From: knepley at gmail.com (Matthew Knepley) Date: Fri, 12 Oct 2007 02:59:01 -0500 Subject: configure stalling on MacOSX In-Reply-To: <20071012072820.1D5D16AE@bonito.ulb.ac.be> References: <20071012072820.1D5D16AE@bonito.ulb.ac.be> Message-ID: On 10/12/07, Knaepen Bernard wrote: > Hello, > > is it normal that the ./config/configure.py stalls for a long time when building petsc-2.3.3-p6 on MacOSX? The step at which it stays stuck is: > > TESTING: checkSharedLinkerPaths from config.setCompilers(python/BuildSystem/config/setCompilers.py:1107) > > It remains there for around five minutes then continues the configure process without problem. Except that it leaves a fortran process alive that I have to manually kill later on. This process looks like: The 5 minutes is a timeout value. This test has to run an executable and check for output. It seems that the run never finishes. However, after timeout it should kill the original thread. Thus I am guessing that this is a problem with the Python on OSX, however it could also be my code. Matt > 17623 fortcom 60.8% 5:06.44 1 13 60 364K 12.1M 2.23M 57.8M > 17614 ifortia32 0.0% 0:00.02 1 14 19 392K 1.96M 1.86M 29.0M > 17613 ifortbin 0.0% 0:00.01 1 13 16 148K 992K 876K 28.6M > > when I use the top command. > > I am configuring with the following options: > > /config/configure.py --with-cc=mpicc --with-fc=mpif90 --with-hypre=1 --download-hypre=1 --with-debugging=0 COPTFLAGS=-O3 FOPTFLAGS=-O3 --with-cxx=mpicxx --with-x=0 --with-sudo=sudo > > The system is MacOSX 10.4.9 (latest) running on a MacPro (or Macbook Pro) and the fortran compiler is ifort 10.0.17. > > Regards, > Bernard. > > > > -- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. -- Norbert Wiener From dimitri.lecas at c-s.fr Fri Oct 12 07:31:59 2007 From: dimitri.lecas at c-s.fr (LECAS Dimitri) Date: Fri, 12 Oct 2007 14:31:59 +0200 Subject: Solve a symmetric matrix Message-ID: <2ad9428eb8.28eb82ad94@c-s.fr> ----- Original Message ----- From: Hong Zhang Date: Tuesday, October 2, 2007 3:51 pm Subject: Re: Solve a symmetric matrix > > PCILU and PCICC are not supported for MPISBAIJ matrix type > (the error message should indicate this problem). > For sequential sbaij matrix, you can use PCICC. > For parallel, you may try the default petsc preconditioner > block jacobi + icc. See Chapter 4 of petsc user manual. > Ok i see. The problem came from the call to KSPSetUp, when using one MPI task, this call end in MatILUFactorSymbolic with the error message about symbolic ILU with Matrix type seqsbaij. With more than one MPI task, there is no more error. -- Dimitri Lecas From hzhang at mcs.anl.gov Fri Oct 12 09:06:43 2007 From: hzhang at mcs.anl.gov (Hong Zhang) Date: Fri, 12 Oct 2007 09:06:43 -0500 (CDT) Subject: Solve a symmetric matrix In-Reply-To: <2ad9428eb8.28eb82ad94@c-s.fr> References: <2ad9428eb8.28eb82ad94@c-s.fr> Message-ID: > > PCILU and PCICC are not supported for MPISBAIJ matrix type > > (the error message should indicate this problem). > > For sequential sbaij matrix, you can use PCICC. > > For parallel, you may try the default petsc preconditioner > > block jacobi + icc. See Chapter 4 of petsc user manual. > > > > Ok i see. > > The problem came from the call to KSPSetUp, when using one MPI task, > this call end in MatILUFactorSymbolic with the error message about > symbolic ILU with Matrix type seqsbaij. > > With more than one MPI task, there is no more error. I would suggest you using '-ksp_view' to check which default solvers are being used for both cases. I guess the "block jacobi + icc" is used on parallel implementation. Hong From balay at mcs.anl.gov Fri Oct 12 09:23:18 2007 From: balay at mcs.anl.gov (Satish Balay) Date: Fri, 12 Oct 2007 09:23:18 -0500 (CDT) Subject: Version of Mpi and fblaslapack In-Reply-To: References: Message-ID: On Fri, 12 Oct 2007, Mads Hoel wrote: > > I found out that linking PETSc with another library (Diffpack) compiled with > different > versions of MPI will crash when calling InitializePetsc(...). My solution was > to drop running config/configure with --download-mpi option, but have MPI > point to the correct installation of MPI using the --with-mpi-dir=/path/to/mpi > option. Yes - you cannot mix 2 different versions of MPI [or any other package] to create a single binary. This is equvalent to using include files from package-version1 - and linking with package-version2. [which is obviously wrong.] In the same way, its best to install ALL packages that go into a binary with the same compiler set [c,c++,fortran]. Usually compilers & MPI are the basic blocks that many external packages use, so in this case one can first install MPI [with the correct desired compilers]. And then install all other packages [used by the application] with the MPI wrappers. And all external packages built by PETSc will use the compilers/MPI specified to PETSc. ./config/configure.py --with-cc=mpicc --with-fc=mpif90 or the shortcut equivalent of the above: ./config/configure.py --with-mpi-dir= Satish From jagruti.trivedi at navy.mil Fri Oct 12 12:05:01 2007 From: jagruti.trivedi at navy.mil (Trivedi, Jagruti CIV 470000D, 474300D) Date: Fri, 12 Oct 2007 10:05:01 -0700 Subject: Version of Mpi and fblaslapack In-Reply-To: References: Message-ID: Satish, That is not the point. The question is how to find out which version of compilers such as gcc(or c ) , C++ mpicc , mpif77 is needed for petsc 2.2.1, 2.3.0, 2.3.2 etc? I do not want to find out after the fact that oh I needed so and so compiler version. If possible, can this information be listed as part of Release Notes? If this information is already available, can you point me to the right place. Thanks, ...Jagruti -----Original Message----- From: owner-petsc-users at mcs.anl.gov [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Satish Balay Sent: Friday, October 12, 2007 7:23 To: petsc-users at mcs.anl.gov Subject: Re: Version of Mpi and fblaslapack On Fri, 12 Oct 2007, Mads Hoel wrote: > > I found out that linking PETSc with another library (Diffpack) > compiled with different versions of MPI will crash when calling > InitializePetsc(...). My solution was to drop running config/configure > with --download-mpi option, but have MPI point to the correct > installation of MPI using the --with-mpi-dir=/path/to/mpi option. Yes - you cannot mix 2 different versions of MPI [or any other package] to create a single binary. This is equvalent to using include files from package-version1 - and linking with package-version2. [which is obviously wrong.] In the same way, its best to install ALL packages that go into a binary with the same compiler set [c,c++,fortran]. Usually compilers & MPI are the basic blocks that many external packages use, so in this case one can first install MPI [with the correct desired compilers]. And then install all other packages [used by the application] with the MPI wrappers. And all external packages built by PETSc will use the compilers/MPI specified to PETSc. ./config/configure.py --with-cc=mpicc --with-fc=mpif90 or the shortcut equivalent of the above: ./config/configure.py --with-mpi-dir= Satish From dalcinl at gmail.com Fri Oct 12 12:09:41 2007 From: dalcinl at gmail.com (Lisandro Dalcin) Date: Fri, 12 Oct 2007 14:09:41 -0300 Subject: Preconditioner for bicg In-Reply-To: References: <925346A443D4E340BEB20248BAFCDBDF02A11CEC@CFEVS1-IP.americas.cray.com> <925346A443D4E340BEB20248BAFCDBDF02A11D23@CFEVS1-IP.americas.cray.com> Message-ID: But for cgne, looking at the code, ApplyTranspose seems to be optional... On 10/11/07, Barry Smith wrote: > > [bsmith:~] bsmith% cd Src/petsc-dev/src/ksp/ksp/impls > [bsmith:ksp/ksp/impls] bsmith% grep ApplyTranspose */*.c > bicg/bicg.c: ierr = KSP_PCApplyTranspose(ksp,Rl,Zl);CHKERRQ(ierr); > bicg/bicg.c: ierr = KSP_PCApplyTranspose(ksp,Rl,Zl);CHKERRQ(ierr); > bicg/bicg.c: ierr = KSP_PCApplyTranspose(ksp,Rl,Zl);CHKERRQ(ierr); > > [bsmith:ksp/ksp/impls] bsmith% grep ApplyTranspose */*/*.c > cg/cgne/cgne.c: ierr = PCApplyTransposeExists(ksp->pc,&transpose_pc);CHKERRQ(ierr); > cg/cgne/cgne.c: ierr = KSP_PCApplyTranspose(ksp,T,Z);CHKERRQ(ierr); > cg/cgne/cgne.c: ierr = KSP_PCApplyTranspose(ksp,T,Z);CHKERRQ(ierr); > > These are the two methods that require the transpose operation. > > I have documented both of their manual pages to indicate this. > > Barry > > > On Thu, 11 Oct 2007, Keita Teranishi wrote: > > > Hi Barry, > > > > Could you suggest any other KSP solvers that have the same type of problems? > > > > Thanks, > > ================================ > > Keita Teranishi > > Math Software Group > > Cray Inc. > > keita at cray.com > > ================================ > > > > -----Original Message----- > > From: owner-petsc-users at mcs.anl.gov [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Barry Smith > > Sent: Thursday, October 11, 2007 4:50 PM > > To: petsc-users at mcs.anl.gov > > Subject: Re: Preconditioner for bicg > > > > > > For some preconditioners it is simply not easily possible to apply the > > transpose of the preconditioner. Even multigrid requires some work to do this. > > > > Barry > > > > > > On Thu, 11 Oct 2007, Keita Teranishi wrote: > > > > > Hi, > > > > > > > > > > > > I'd like to make sure preconditioners supported for bicg. I found bicg only accepts preconditioners that is interfaced through PCApplyTranspose. > > > > > > For this reason, I see several preconditioners fails with bicg. > > > > > > > > > > > > Thanks, > > > > > > ================================ > > > Keita Teranishi > > > Math Software Group > > > Cray Inc. > > > keita at cray.com > > > ================================ > > > > > > > > > > > > > > > > > > -- Lisandro Dalc?n --------------- Centro Internacional de M?todos Computacionales en Ingenier?a (CIMEC) Instituto de Desarrollo Tecnol?gico para la Industria Qu?mica (INTEC) Consejo Nacional de Investigaciones Cient?ficas y T?cnicas (CONICET) PTLC - G?emes 3450, (3000) Santa Fe, Argentina Tel/Fax: +54-(0)342-451.1594 From bsmith at mcs.anl.gov Fri Oct 12 12:16:48 2007 From: bsmith at mcs.anl.gov (Barry Smith) Date: Fri, 12 Oct 2007 12:16:48 -0500 (CDT) Subject: Preconditioner for bicg In-Reply-To: References: <925346A443D4E340BEB20248BAFCDBDF02A11CEC@CFEVS1-IP.americas.cray.com> <925346A443D4E340BEB20248BAFCDBDF02A11D23@CFEVS1-IP.americas.cray.com> Message-ID: You are right. What was Victor thinking? Seems a pretty strange algorithm in that case. I have updated the manual page yet again to reflect this. Barry I am not sure we really want to have this behavior, seems too likely to perform badly. On Fri, 12 Oct 2007, Lisandro Dalcin wrote: > But for cgne, looking at the code, ApplyTranspose seems to be optional... > > On 10/11/07, Barry Smith wrote: > > > > [bsmith:~] bsmith% cd Src/petsc-dev/src/ksp/ksp/impls > > [bsmith:ksp/ksp/impls] bsmith% grep ApplyTranspose */*.c > > bicg/bicg.c: ierr = KSP_PCApplyTranspose(ksp,Rl,Zl);CHKERRQ(ierr); > > bicg/bicg.c: ierr = KSP_PCApplyTranspose(ksp,Rl,Zl);CHKERRQ(ierr); > > bicg/bicg.c: ierr = KSP_PCApplyTranspose(ksp,Rl,Zl);CHKERRQ(ierr); > > > > [bsmith:ksp/ksp/impls] bsmith% grep ApplyTranspose */*/*.c > > cg/cgne/cgne.c: ierr = PCApplyTransposeExists(ksp->pc,&transpose_pc);CHKERRQ(ierr); > > cg/cgne/cgne.c: ierr = KSP_PCApplyTranspose(ksp,T,Z);CHKERRQ(ierr); > > cg/cgne/cgne.c: ierr = KSP_PCApplyTranspose(ksp,T,Z);CHKERRQ(ierr); > > > > These are the two methods that require the transpose operation. > > > > I have documented both of their manual pages to indicate this. > > > > Barry > > > > > > On Thu, 11 Oct 2007, Keita Teranishi wrote: > > > > > Hi Barry, > > > > > > Could you suggest any other KSP solvers that have the same type of problems? > > > > > > Thanks, > > > ================================ > > > Keita Teranishi > > > Math Software Group > > > Cray Inc. > > > keita at cray.com > > > ================================ > > > > > > -----Original Message----- > > > From: owner-petsc-users at mcs.anl.gov [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Barry Smith > > > Sent: Thursday, October 11, 2007 4:50 PM > > > To: petsc-users at mcs.anl.gov > > > Subject: Re: Preconditioner for bicg > > > > > > > > > For some preconditioners it is simply not easily possible to apply the > > > transpose of the preconditioner. Even multigrid requires some work to do this. > > > > > > Barry > > > > > > > > > On Thu, 11 Oct 2007, Keita Teranishi wrote: > > > > > > > Hi, > > > > > > > > > > > > > > > > I'd like to make sure preconditioners supported for bicg. I found bicg only accepts preconditioners that is interfaced through PCApplyTranspose. > > > > > > > > For this reason, I see several preconditioners fails with bicg. > > > > > > > > > > > > > > > > Thanks, > > > > > > > > ================================ > > > > Keita Teranishi > > > > Math Software Group > > > > Cray Inc. > > > > keita at cray.com > > > > ================================ > > > > > > > > > > > > > > > > > > > > > > > > > > > > > From bsmith at mcs.anl.gov Fri Oct 12 12:41:32 2007 From: bsmith at mcs.anl.gov (Barry Smith) Date: Fri, 12 Oct 2007 12:41:32 -0500 (CDT) Subject: Version of Mpi and fblaslapack In-Reply-To: References: Message-ID: You are totally misunderstanding the situation! PETSc should work with virtually ANY release of gcc or other C or C++ compiler and ANY mpicc and mpif77 unless they have bugs or are not installed properly. The point is you can not mix incompatible compilers or MPIs when you install two or more different packages to work together. This has nothing to do with PETSc. If you build Diffpack and Superlu_dist with different MPI's it would not work. Barry On Fri, 12 Oct 2007, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > Satish, > > That is not the point. The question is how to find out which version of > compilers such as gcc(or c ) , C++ mpicc , mpif77 is needed for petsc > 2.2.1, 2.3.0, 2.3.2 etc? > > I do not want to find out after the fact that oh I needed so and so > compiler version. > > If possible, can this information be listed as part of Release Notes? If > this information is already available, can you point me to the right > place. > > Thanks, > > ...Jagruti > -----Original Message----- > From: owner-petsc-users at mcs.anl.gov > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Satish Balay > Sent: Friday, October 12, 2007 7:23 > To: petsc-users at mcs.anl.gov > Subject: Re: Version of Mpi and fblaslapack > > On Fri, 12 Oct 2007, Mads Hoel wrote: > > > > I found out that linking PETSc with another library (Diffpack) > > compiled with different versions of MPI will crash when calling > > InitializePetsc(...). My solution was to drop running config/configure > > > with --download-mpi option, but have MPI point to the correct > > installation of MPI using the --with-mpi-dir=/path/to/mpi option. > > Yes - you cannot mix 2 different versions of MPI [or any other package] > to create a single binary. This is equvalent to using include files from > package-version1 - and linking with package-version2. [which is > obviously wrong.] > > In the same way, its best to install ALL packages that go into a binary > with the same compiler set [c,c++,fortran]. > > Usually compilers & MPI are the basic blocks that many external packages > use, so in this case one can first install MPI [with the correct desired > compilers]. And then install all other packages [used by the > application] with the MPI wrappers. > > And all external packages built by PETSc will use the compilers/MPI > specified to PETSc. > > ./config/configure.py --with-cc=mpicc --with-fc=mpif90 or the shortcut > equivalent of the above: > ./config/configure.py --with-mpi-dir= > > Satish > > From jagruti.trivedi at navy.mil Fri Oct 12 12:59:18 2007 From: jagruti.trivedi at navy.mil (Trivedi, Jagruti CIV 470000D, 474300D) Date: Fri, 12 Oct 2007 10:59:18 -0700 Subject: Version of Mpi and fblaslapack In-Reply-To: References: Message-ID: Yes, I am sorry. I did misunderstood the situation. Best Regards, ...Jagruti -----Original Message----- From: owner-petsc-users at mcs.anl.gov [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Barry Smith Sent: Friday, October 12, 2007 10:42 To: petsc-users at mcs.anl.gov Subject: RE: Version of Mpi and fblaslapack You are totally misunderstanding the situation! PETSc should work with virtually ANY release of gcc or other C or C++ compiler and ANY mpicc and mpif77 unless they have bugs or are not installed properly. The point is you can not mix incompatible compilers or MPIs when you install two or more different packages to work together. This has nothing to do with PETSc. If you build Diffpack and Superlu_dist with different MPI's it would not work. Barry On Fri, 12 Oct 2007, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > Satish, > > That is not the point. The question is how to find out which version > of compilers such as gcc(or c ) , C++ mpicc , mpif77 is needed for > petsc 2.2.1, 2.3.0, 2.3.2 etc? > > I do not want to find out after the fact that oh I needed so and so > compiler version. > > If possible, can this information be listed as part of Release Notes? > If this information is already available, can you point me to the > right place. > > Thanks, > > ...Jagruti > -----Original Message----- > From: owner-petsc-users at mcs.anl.gov > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Satish Balay > Sent: Friday, October 12, 2007 7:23 > To: petsc-users at mcs.anl.gov > Subject: Re: Version of Mpi and fblaslapack > > On Fri, 12 Oct 2007, Mads Hoel wrote: > > > > I found out that linking PETSc with another library (Diffpack) > > compiled with different versions of MPI will crash when calling > > InitializePetsc(...). My solution was to drop running > > config/configure > > > with --download-mpi option, but have MPI point to the correct > > installation of MPI using the --with-mpi-dir=/path/to/mpi option. > > Yes - you cannot mix 2 different versions of MPI [or any other > package] to create a single binary. This is equvalent to using include > files from > package-version1 - and linking with package-version2. [which is > obviously wrong.] > > In the same way, its best to install ALL packages that go into a > binary with the same compiler set [c,c++,fortran]. > > Usually compilers & MPI are the basic blocks that many external > packages use, so in this case one can first install MPI [with the > correct desired compilers]. And then install all other packages [used > by the application] with the MPI wrappers. > > And all external packages built by PETSc will use the compilers/MPI > specified to PETSc. > > ./config/configure.py --with-cc=mpicc --with-fc=mpif90 or the shortcut > equivalent of the above: > ./config/configure.py --with-mpi-dir= > > Satish > > From balay at mcs.anl.gov Fri Oct 12 13:10:05 2007 From: balay at mcs.anl.gov (Satish Balay) Date: Fri, 12 Oct 2007 13:10:05 -0500 (CDT) Subject: Version of Mpi and fblaslapack In-Reply-To: References: Message-ID: petsc-2.2.1 was released in Aug2004 - hence tested with compilers available at that time - whereas petsc-2.3.3 is tested with compilers available to us as of May2007 [some old compilers, some new compilers] So it is possible that there are some issues with certain combinations. We don't have a document that lists these compatibilities or incompatibilities. If you encounter problems - let us know at petsc-maint at mcs.anl.gov - we can help you resolve them. We generally hope that users migrate to newer versions of PETSc, so that we don't have to deal with such issues that come up with aging old versions of software.. Satish On Fri, 12 Oct 2007, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > > Yes, I am sorry. I did misunderstood the situation. > > Best Regards, > > ...Jagruti > -----Original Message----- > From: owner-petsc-users at mcs.anl.gov > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Barry Smith > Sent: Friday, October 12, 2007 10:42 > To: petsc-users at mcs.anl.gov > Subject: RE: Version of Mpi and fblaslapack > > > You are totally misunderstanding the situation! PETSc should work with > virtually ANY release of gcc or other C or C++ compiler and ANY mpicc > and mpif77 unless they have bugs or are not installed properly. > > The point is you can not mix incompatible compilers or MPIs when you > install two or more different packages to work together. This has > nothing to do with PETSc. If you build Diffpack and Superlu_dist with > different MPI's it would not work. > > Barry > > > On Fri, 12 Oct 2007, Trivedi, Jagruti CIV 470000D, 474300D wrote: > > > > > Satish, > > > > That is not the point. The question is how to find out which version > > of compilers such as gcc(or c ) , C++ mpicc , mpif77 is needed for > > petsc 2.2.1, 2.3.0, 2.3.2 etc? > > > > I do not want to find out after the fact that oh I needed so and so > > compiler version. > > > > If possible, can this information be listed as part of Release Notes? > > If this information is already available, can you point me to the > > right place. > > > > Thanks, > > > > ...Jagruti > > -----Original Message----- > > From: owner-petsc-users at mcs.anl.gov > > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Satish Balay > > Sent: Friday, October 12, 2007 7:23 > > To: petsc-users at mcs.anl.gov > > Subject: Re: Version of Mpi and fblaslapack > > > > On Fri, 12 Oct 2007, Mads Hoel wrote: > > > > > > I found out that linking PETSc with another library (Diffpack) > > > compiled with different versions of MPI will crash when calling > > > InitializePetsc(...). My solution was to drop running > > > config/configure > > > > > with --download-mpi option, but have MPI point to the correct > > > installation of MPI using the --with-mpi-dir=/path/to/mpi option. > > > > Yes - you cannot mix 2 different versions of MPI [or any other > > package] to create a single binary. This is equvalent to using include > > > files from > > package-version1 - and linking with package-version2. [which is > > obviously wrong.] > > > > In the same way, its best to install ALL packages that go into a > > binary with the same compiler set [c,c++,fortran]. > > > > Usually compilers & MPI are the basic blocks that many external > > packages use, so in this case one can first install MPI [with the > > correct desired compilers]. And then install all other packages [used > > by the application] with the MPI wrappers. > > > > And all external packages built by PETSc will use the compilers/MPI > > specified to PETSc. > > > > ./config/configure.py --with-cc=mpicc --with-fc=mpif90 or the shortcut > > > equivalent of the above: > > ./config/configure.py --with-mpi-dir= > > > > Satish > > > > > > From petsc-maint at mcs.anl.gov Fri Oct 12 13:54:00 2007 From: petsc-maint at mcs.anl.gov (Barry Smith) Date: Fri, 12 Oct 2007 13:54:00 -0500 (CDT) Subject: references for teragrid proposal In-Reply-To: <6B280ADC-0954-443F-A600-6B8CD2899610@ldeo.columbia.edu> References: <6B280ADC-0954-443F-A600-6B8CD2899610@ldeo.columbia.edu> Message-ID: On Thu, 11 Oct 2007, Samar Khatiwala wrote: > Hello > > I am writing a proposal for a largish amount of compute time allocation on the > NSF teragrid > (http://www.ci-partnership.org/Allocations/). This is for a "Medium Resource > Allocation" (MRAC) > in case you are familiar with the procedure. > > My proposed research uses my tracer advection-diffusion code based entirely on > PETSc, and I want to make > the case that the teragrid machines are well suited to running this software. > Can someone please suggest > references I can cite that will help me make a strong case? (FYI: Sparse > mat-vec is probably the most > often used operation in my code.) > > The machines I plan on using are: > > TACC PowerEdge 1955 (Lonestar) : Xeon, EM64T, dual CPU (dual-core) 2.6 GHz, > InfiniBand Switch > NCSA IA-64 Cluster (Mercury): Itanium2, 1.5 GHz, Myrinet > NCSA IA-64 Linux Cluster (Abe): Intel 64, 2.3 GHz, InfiniBand > SDSC IA-64 Cluster: Itanium2, 1.5 GHz, Myrinet > > (Hardware details are here: > http://www.teragrid.org/userinfo/hardware/resources.php) > > Thanks very much > > Samar > > > From bsmith at mcs.anl.gov Fri Oct 12 16:26:07 2007 From: bsmith at mcs.anl.gov (Barry Smith) Date: Fri, 12 Oct 2007 16:26:07 -0500 (CDT) Subject: configure stalling on MacOSX In-Reply-To: <20071012072820.1D5D16AE@bonito.ulb.ac.be> References: <20071012072820.1D5D16AE@bonito.ulb.ac.be> Message-ID: I have run with 10.0.0.20 on the Mac with no difficulties using the options below with petsc-2.3.3-6 and have no hanging. config/configure.py just takes a couple of minutes. [bsmith:~/Src/petsc-release-2.3.3] bsmith% uname -a Darwin bsmith.mcs.anl.gov 8.10.1 Darwin Kernel Version 8.10.1: Wed May 23 16:33:00 PDT 2007; root:xnu-792.22.5~1/RELEASE_I386 i386 i386 Barry On Fri, 12 Oct 2007, Knaepen Bernard wrote: > Hello, > > is it normal that the ./config/configure.py stalls for a long time when building petsc-2.3.3-p6 on MacOSX? The step at which it stays stuck is: > > TESTING: checkSharedLinkerPaths from config.setCompilers(python/BuildSystem/config/setCompilers.py:1107) > > It remains there for around five minutes then continues the configure process without problem. Except that it leaves a fortran process alive that I have to manually kill later on. This process looks like: > > 17623 fortcom 60.8% 5:06.44 1 13 60 364K 12.1M 2.23M 57.8M > 17614 ifortia32 0.0% 0:00.02 1 14 19 392K 1.96M 1.86M 29.0M > 17613 ifortbin 0.0% 0:00.01 1 13 16 148K 992K 876K 28.6M > > when I use the top command. > > I am configuring with the following options: > > /config/configure.py --with-cc=mpicc --with-fc=mpif90 --with-hypre=1 --download-hypre=1 --with-debugging=0 COPTFLAGS=-O3 FOPTFLAGS=-O3 --with-cxx=mpicxx --with-x=0 --with-sudo=sudo > > The system is MacOSX 10.4.9 (latest) running on a MacPro (or Macbook Pro) and the fortran compiler is ifort 10.0.17. > > Regards, > Bernard. > > > > From bknaepen at ulb.ac.be Sat Oct 13 11:03:14 2007 From: bknaepen at ulb.ac.be (Bernard Knaepen) Date: Sat, 13 Oct 2007 18:03:14 +0200 Subject: configure stalling on MacOSX In-Reply-To: References: <20071012072820.1D5D16AE@bonito.ulb.ac.be> Message-ID: Hi Matthew and Barry, I tried again after installing intel fortran 10.0.020 but I still get the hanging. I copy below the output from configure.log that exhibits the timeout. When I remove the test related to '-rpath ' then the scripts does not hang. No problems either if I disable fortran with -- with-fc=0 uname -a gives: Darwin dolfin.local 8.10.1 Darwin Kernel Version 8.10.1: Wed May 23 16:33:00 PDT 2007; root:xnu-792.22.5~1/RELEASE_I386 i386 i386 Regards, Bernard. ======================================================================== ======== TEST checkSharedLinkerPaths from config.setCompilers(/Users/bknaepen/ Desktop/petsc-2.3.3-p6/python/BuildSystem/config/setCompilers.py:1107) TESTING: checkSharedLinkerPaths from config.setCompilers(python/ BuildSystem/config/setCompilers.py:1107) Determine the shared linker path options - IRIX: -rpath - Linux, OSF: -Wl,-rpath, - Solaris: -R - FreeBSD: -Wl,-R, Pushing language C sh: mpicc -flags Executing: mpicc -flags sh: Trying C linker flag -Wl,-rpath, sh: mpicc -c -o conftest.o -fPIC conftest.c Executing: mpicc -c -o conftest.o -fPIC conftest.c sh: Pushing language C Popping language C sh: mpicc -o conftest -Wl,-rpath,/Users/bknaepen/Desktop/ petsc-2.3.3-p6 -Wl,-multiply_defined,suppress -Wl,-multiply_defined - Wl,suppress -Wl,-multiply_defined,suppress -Wl,-multiply_defined - Wl,suppress -fPIC conftest.o Executing: mpicc -o conftest -Wl,-rpath,/Users/bknaepen/Desktop/ petsc-2.3.3-p6 -Wl,-multiply_defined,suppress -Wl,-multiply_defined - Wl,suppress -Wl,-multiply_defined,suppress -Wl,-multiply_defined - Wl,suppress -fPIC conftest.o sh: Possible ERROR while running linker: /usr/bin/ld: unknown flag: -rpath collect2: ld returned 1 exit status output: ret = 256 error message = {/usr/bin/ld: unknown flag: -rpath collect2: ld returned 1 exit status } Pushing language C Popping language C in mpicc -o conftest -Wl,-rpath,/Users/bknaepen/Desktop/ petsc-2.3.3-p6 -Wl,-multiply_defined,suppress -Wl,-multiply_defined - Wl,suppress -Wl,-multiply_defined,suppress -Wl,-multiply_defined - Wl,suppress -fPIC conftest.o Source: #include "confdefs.h" #include "conffix.h" int main() { ; return 0; } Rejecting linker flag -Wl,-rpath,/Users/bknaepen/Desktop/ petsc-2.3.3-p6 due to nonzero status from link Rejecting C linker flag -Wl,-rpath,/Users/bknaepen/Desktop/ petsc-2.3.3-p6 due to /usr/bin/ld: unknown flag: -rpath collect2: ld returned 1 exit status Rejected C linker flag -Wl,-rpath, Trying C linker flag -R sh: mpicc -c -o conftest.o -fPIC conftest.c Executing: mpicc -c -o conftest.o -fPIC conftest.c sh: Pushing language C Popping language C sh: mpicc -o conftest -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 -Wl,- multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -Wl,- multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC conftest.o Executing: mpicc -o conftest -R/Users/bknaepen/Desktop/petsc-2.3.3- p6 -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress - Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress - fPIC conftest.o sh: Possible ERROR while running linker: i686-apple-darwin8-gcc-4.0.1: unrecognized option '-R/Users/bknaepen/Desktop/petsc-2.3.3-p6' output: error message = {i686-apple-darwin8-gcc-4.0.1: unrecognized option '-R/Users/bknaepen/Desktop/petsc-2.3.3-p6' } Pushing language C Popping language C in mpicc -o conftest -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 - Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -Wl,- multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC conftest.o Source: #include "confdefs.h" #include "conffix.h" int main() { ; return 0; } Rejecting C linker flag -R/Users/bknaepen/Desktop/ petsc-2.3.3-p6 due to i686-apple-darwin8-gcc-4.0.1: unrecognized option '-R/Users/bknaepen/ Desktop/petsc-2.3.3-p6' Rejected C linker flag -R Trying C linker flag -rpath sh: mpicc -c -o conftest.o -fPIC conftest.c Executing: mpicc -c -o conftest.o -fPIC conftest.c sh: Pushing language C Popping language C sh: mpicc -o conftest -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress - Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress - fPIC conftest.o Executing: mpicc -o conftest -rpath /Users/bknaepen/Desktop/ petsc-2.3.3-p6 -Wl,-multiply_defined,suppress -Wl,-multiply_defined - Wl,suppress -Wl,-multiply_defined,suppress -Wl,-multiply_defined - Wl,suppress -fPIC conftest.o sh: Possible ERROR while running linker: i686-apple-darwin8-gcc-4.0.1: unrecognized option '-rpath' /usr/bin/ld: can't map file: /Users/bknaepen/Desktop/petsc-2.3.3-p6 ((os/kern) invalid argument) collect2: ld returned 1 exit status output: ret = 256 error message = {i686-apple-darwin8-gcc-4.0.1: unrecognized option '- rpath' /usr/bin/ld: can't map file: /Users/bknaepen/Desktop/petsc-2.3.3-p6 ((os/kern) invalid argument) collect2: ld returned 1 exit status } Pushing language C Popping language C in mpicc -o conftest -rpath /Users/bknaepen/Desktop/petsc-2.3.3- p6 -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress - Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress - fPIC conftest.o Source: #include "confdefs.h" #include "conffix.h" int main() { ; return 0; } Rejecting linker flag -rpath /Users/bknaepen/Desktop/ petsc-2.3.3-p6 due to nonzero status from link Rejecting C linker flag -rpath /Users/bknaepen/Desktop/ petsc-2.3.3-p6 due to i686-apple-darwin8-gcc-4.0.1: unrecognized option '-rpath' /usr/bin/ld: can't map file: /Users/bknaepen/Desktop/petsc-2.3.3-p6 ((os/kern) invalid argument) collect2: ld returned 1 exit status Rejected C linker flag -rpath Trying C linker flag -Wl,-R, sh: mpicc -c -o conftest.o -fPIC conftest.c Executing: mpicc -c -o conftest.o -fPIC conftest.c sh: Pushing language C Popping language C sh: mpicc -o conftest -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress - Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress - fPIC conftest.o Executing: mpicc -o conftest -Wl,-R,/Users/bknaepen/Desktop/ petsc-2.3.3-p6 -Wl,-multiply_defined,suppress -Wl,-multiply_defined - Wl,suppress -Wl,-multiply_defined,suppress -Wl,-multiply_defined - Wl,suppress -fPIC conftest.o sh: Possible ERROR while running linker: /usr/bin/ld: unknown flag: -R collect2: ld returned 1 exit status output: ret = 256 error message = {/usr/bin/ld: unknown flag: -R collect2: ld returned 1 exit status } Pushing language C Popping language C in mpicc -o conftest -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3- p6 -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress - Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress - fPIC conftest.o Source: #include "confdefs.h" #include "conffix.h" int main() { ; return 0; } Rejecting linker flag -Wl,-R,/Users/bknaepen/Desktop/ petsc-2.3.3-p6 due to nonzero status from link Rejecting C linker flag -Wl,-R,/Users/bknaepen/Desktop/ petsc-2.3.3-p6 due to /usr/bin/ld: unknown flag: -R collect2: ld returned 1 exit status Rejected C linker flag -Wl,-R, Popping language C Pushing language FC sh: mpicc -flags Executing: mpicc -flags sh: Trying FC linker flag -Wl,-rpath, sh: mpif90 -c -o conftest.o -I. -fPIC conftest.F Executing: mpif90 -c -o conftest.o -I. -fPIC conftest.F sh: Pushing language C Popping language C sh: mpif90 -o conftest -Wl,-rpath,/Users/bknaepen/Desktop/ petsc-2.3.3-p6 -I. -fPIC conftest.o Executing: mpif90 -o conftest -Wl,-rpath,/Users/bknaepen/Desktop/ petsc-2.3.3-p6 -I. -fPIC conftest.o sh: Possible ERROR while running linker: ipo: error #11019: OBJREAD Error: Could not create mapping for /Users/bknaepen/Desktop/ petsc-2.3.3-p6 ifort: error #10014: problem during multi-file optimization compilation (code 1) output: ret = 256 error message = {ipo: error #11019: OBJREAD Error: Could not create mapping for /Users/bknaepen/Desktop/petsc-2.3.3-p6 ifort: error #10014: problem during multi-file optimization compilation (code 1) } Pushing language C Popping language C in mpif90 -o conftest -Wl,-rpath,/Users/bknaepen/Desktop/ petsc-2.3.3-p6 -I. -fPIC conftest.o Source: program main end Rejecting linker flag -Wl,-rpath,/Users/bknaepen/Desktop/ petsc-2.3.3-p6 due to nonzero status from link Valid FC linker flag -Wl,-rpath,/Users/bknaepen/Desktop/ petsc-2.3.3-p6 Rejected FC linker flag -Wl,-rpath, Trying FC linker flag -R sh: mpif90 -c -o conftest.o -I. -fPIC conftest.F Executing: mpif90 -c -o conftest.o -I. -fPIC conftest.F sh: Pushing language C Popping language C sh: mpif90 -o conftest -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 - I. -fPIC conftest.o Executing: mpif90 -o conftest -R/Users/bknaepen/Desktop/petsc-2.3.3- p6 -I. -fPIC conftest.o sh: Possible ERROR while running linker: ifort: command line warning #10006: ignoring unknown option '-R/Users/bknaepen/Desktop/ petsc-2.3.3-p6' output: error message = {ifort: command line warning #10006: ignoring unknown option '-R/Users/bknaepen/Desktop/petsc-2.3.3-p6' } Pushing language C Popping language C in mpif90 -o conftest -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 - I. -fPIC conftest.o Source: program main end Rejecting FC linker flag -R/Users/bknaepen/Desktop/ petsc-2.3.3-p6 due to ifort: command line warning #10006: ignoring unknown option '-R/Users/ bknaepen/Desktop/petsc-2.3.3-p6' Rejected FC linker flag -R Trying FC linker flag -rpath sh: mpif90 -c -o conftest.o -I. -fPIC conftest.F Executing: mpif90 -c -o conftest.o -I. -fPIC conftest.F sh: Pushing language C Popping language C sh: mpif90 -o conftest -rpath /Users/bknaepen/Desktop/petsc-2.3.3- p6 -I. -fPIC conftest.o Executing: mpif90 -o conftest -rpath /Users/bknaepen/Desktop/ petsc-2.3.3-p6 -I. -fPIC conftest.o Runaway process exceeded time limit of 600.0s Possible ERROR while running linker: Runaway process exceeded time limit of 600.0s output: Runaway processret = -1 error message = {Runaway process exceeded time limit of 600.0s } Pushing language C Popping language C in mpif90 -o conftest -rpath /Users/bknaepen/Desktop/petsc-2.3.3- p6 -I. -fPIC conftest.o Source: program main end Rejecting linker flag -rpath /Users/bknaepen/Desktop/ petsc-2.3.3-p6 due to nonzero status from link Valid FC linker flag -rpath /Users/bknaepen/Desktop/ petsc-2.3.3-p6 Rejected FC linker flag -rpath Trying FC linker flag -Wl,-R, sh: mpif90 -c -o conftest.o -I. -fPIC conftest.F Executing: mpif90 -c -o conftest.o -I. -fPIC conftest.F sh: Pushing language C Popping language C sh: mpif90 -o conftest -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3- p6 -I. -fPIC conftest.o Executing: mpif90 -o conftest -Wl,-R,/Users/bknaepen/Desktop/ petsc-2.3.3-p6 -I. -fPIC conftest.o sh: Possible ERROR while running linker: IPO link: Warning unknown option '-R'. ipo: error #11019: OBJREAD Error: Could not create mapping for /Users/ bknaepen/Desktop/petsc-2.3.3-p6 ifort: error #10014: problem during multi-file optimization compilation (code 1) output: ret = 256 error message = {IPO link: Warning unknown option '-R'. ipo: error #11019: OBJREAD Error: Could not create mapping for /Users/ bknaepen/Desktop/petsc-2.3.3-p6 ifort: error #10014: problem during multi-file optimization compilation (code 1) } Pushing language C Popping language C in mpif90 -o conftest -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3- p6 -I. -fPIC conftest.o Source: program main end Rejecting linker flag -Wl,-R,/Users/bknaepen/Desktop/ petsc-2.3.3-p6 due to nonzero status from link Rejecting FC linker flag -Wl,-R,/Users/bknaepen/Desktop/ petsc-2.3.3-p6 due to IPO link: Warning unknown option '-R'. ipo: error #11019: OBJREAD Error: Could not create mapping for /Users/ bknaepen/Desktop/petsc-2.3.3-p6 ifort: error #10014: problem during multi-file optimization compilation (code 1) Rejected FC linker flag -Wl,-R, Popping language FC ======================================================================== ======== On 12 Oct 2007, at 23:26, Barry Smith wrote: > > I have run with 10.0.0.20 on the Mac with no difficulties using > the options > below with petsc-2.3.3-6 and have no hanging. config/configure.py > just takes a > couple of minutes. > > [bsmith:~/Src/petsc-release-2.3.3] bsmith% uname -a > Darwin bsmith.mcs.anl.gov 8.10.1 Darwin Kernel Version 8.10.1: Wed > May 23 > 16:33:00 PDT 2007; root:xnu-792.22.5~1/RELEASE_I386 i386 i386 > > > Barry > > > On Fri, 12 Oct 2007, Knaepen Bernard wrote: > >> Hello, >> >> is it normal that the ./config/configure.py stalls for a long time >> when building petsc-2.3.3-p6 on MacOSX? The step at which it stays >> stuck is: >> >> TESTING: checkSharedLinkerPaths from config.setCompilers(python/ >> BuildSystem/config/setCompilers.py:1107) >> >> It remains there for around five minutes then continues the >> configure process without problem. Except that it leaves a fortran >> process alive that I have to manually kill later on. This process >> looks like: >> >> 17623 fortcom 60.8% 5:06.44 1 13 60 364K 12.1M >> 2.23M 57.8M >> 17614 ifortia32 0.0% 0:00.02 1 14 19 392K 1.96M >> 1.86M 29.0M >> 17613 ifortbin 0.0% 0:00.01 1 13 16 148K 992K >> 876K 28.6M >> >> when I use the top command. >> >> I am configuring with the following options: >> >> /config/configure.py --with-cc=mpicc --with-fc=mpif90 --with- >> hypre=1 --download-hypre=1 --with-debugging=0 COPTFLAGS=-O3 >> FOPTFLAGS=-O3 --with-cxx=mpicxx --with-x=0 --with-sudo=sudo >> >> The system is MacOSX 10.4.9 (latest) running on a MacPro (or >> Macbook Pro) and the fortran compiler is ifort 10.0.17. >> >> Regards, >> Bernard. >> >> >> >> From bsmith at mcs.anl.gov Sat Oct 13 15:13:00 2007 From: bsmith at mcs.anl.gov (Barry Smith) Date: Sat, 13 Oct 2007 15:13:00 -0500 (CDT) Subject: configure stalling on MacOSX In-Reply-To: References: <20071012072820.1D5D16AE@bonito.ulb.ac.be> Message-ID: You get > Executing: mpif90 -o conftest -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. -fPIC conftest.o > Runaway process exceeded time limit of 600.0s > Possible ERROR while running linker: Runaway process exceeded time limit of 600.0s > output: Runaway processret = -1 > error message = {Runaway process exceeded time limit of 600.0s } > Pushing language C > Popping language C I get (not using mpif90, whose is it anyways MPICH1, MPICH2, Lam, OpenMPI?) sh: ifort -o conftest -rpath /Users/bsmith/Src/petsc-dev -fPIC conftest.o Executing: ifort -o conftest -rpath /Users/bsmith/Src/petsc-dev -fPIC conftest.o sh: Possible ERROR while running linker: ifort: command line warning #10156: ignoring option '-r'; no argument required ifort: error #10106: Fatal error in /opt/intel/fc/10.0.020/bin/fortcom, terminated by bus error ifort: error #10014: problem during multi-file optimization compilation (code 1) output: ret = 256 My guess is that your underlying ifort is crashing with a bus error just like mine, but your mpif90 is not properly handling this and so doesn't end and return like it should. Can you try running the compile line that hangs directly from the shell with both ifort directly and then your mpif90; what happens in each case? (Just make a tiny Fortran program in a file). Perhaps yet another reason for hating the absurd "MPI compiler wrappers"; like MPI is SO-so-so special (because it needs to link against a couple of MPI libraries and maybe a couple of system libraries?) that it cannot be be compiled and linked like EVERY OTHER F*&& Unix package ever written but needs its own special compiler wrappers, make me sick. Barry On Sat, 13 Oct 2007, Bernard Knaepen wrote: > Hi Matthew and Barry, > > I tried again after installing intel fortran 10.0.020 but I still get the > hanging. I copy below the output from configure.log that exhibits the timeout. > When I remove the test related to '-rpath ' then the scripts does not hang. No > problems either if I disable fortran with --with-fc=0 > > uname -a gives: > > Darwin dolfin.local 8.10.1 Darwin Kernel Version 8.10.1: Wed May 23 16:33:00 > PDT 2007; root:xnu-792.22.5~1/RELEASE_I386 i386 i386 > > Regards, > Bernard. > > > > ================================================================================ > TEST checkSharedLinkerPaths from > config.setCompilers(/Users/bknaepen/Desktop/petsc-2.3.3-p6/python/BuildSystem/config/setCompilers.py:1107) > TESTING: checkSharedLinkerPaths from > config.setCompilers(python/BuildSystem/config/setCompilers.py:1107) > Determine the shared linker path options > - IRIX: -rpath > - Linux, OSF: -Wl,-rpath, > - Solaris: -R > - FreeBSD: -Wl,-R, > Pushing language C > sh: mpicc -flags > Executing: mpicc -flags > sh: > Trying C linker flag -Wl,-rpath, > sh: mpicc -c -o conftest.o -fPIC conftest.c > Executing: mpicc -c -o conftest.o -fPIC conftest.c > sh: > Pushing language C > Popping language C > sh: mpicc -o conftest -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > conftest.o > Executing: mpicc -o conftest > -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > conftest.o > sh: > Possible ERROR while running linker: /usr/bin/ld: unknown flag: -rpath > collect2: ld returned 1 exit status > output: ret = 256 > error message = {/usr/bin/ld: unknown flag: -rpath > collect2: ld returned 1 exit status > } > Pushing language C > Popping language C > in mpicc -o conftest -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > conftest.o > Source: > #include "confdefs.h" > #include "conffix.h" > > int main() { > ; > return 0; > } > Rejecting linker flag -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > due to nonzero status from link > Rejecting C linker flag > -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 due to > /usr/bin/ld: unknown flag: -rpath > collect2: ld returned 1 exit status > > Rejected C linker flag -Wl,-rpath, > Trying C linker flag -R > sh: mpicc -c -o conftest.o -fPIC conftest.c > Executing: mpicc -c -o conftest.o -fPIC conftest.c > sh: > Pushing language C > Popping language C > sh: mpicc -o conftest -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > conftest.o > Executing: mpicc -o conftest -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > conftest.o > sh: > Possible ERROR while running linker: i686-apple-darwin8-gcc-4.0.1: > unrecognized option '-R/Users/bknaepen/Desktop/petsc-2.3.3-p6' > output: error message = {i686-apple-darwin8-gcc-4.0.1: unrecognized option > '-R/Users/bknaepen/Desktop/petsc-2.3.3-p6' > } > Pushing language C > Popping language C > in mpicc -o conftest -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > conftest.o > Source: > #include "confdefs.h" > #include "conffix.h" > > int main() { > ; > return 0; > } > Rejecting C linker flag -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 due to > i686-apple-darwin8-gcc-4.0.1: unrecognized option > '-R/Users/bknaepen/Desktop/petsc-2.3.3-p6' > > Rejected C linker flag -R > Trying C linker flag -rpath > sh: mpicc -c -o conftest.o -fPIC conftest.c > Executing: mpicc -c -o conftest.o -fPIC conftest.c > sh: > Pushing language C > Popping language C > sh: mpicc -o conftest -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > conftest.o > Executing: mpicc -o conftest -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > conftest.o > sh: > Possible ERROR while running linker: i686-apple-darwin8-gcc-4.0.1: > unrecognized option '-rpath' > /usr/bin/ld: can't map file: /Users/bknaepen/Desktop/petsc-2.3.3-p6 ((os/kern) > invalid argument) > collect2: ld returned 1 exit status > output: ret = 256 > error message = {i686-apple-darwin8-gcc-4.0.1: unrecognized option '-rpath' > /usr/bin/ld: can't map file: /Users/bknaepen/Desktop/petsc-2.3.3-p6 ((os/kern) > invalid argument) > collect2: ld returned 1 exit status > } > Pushing language C > Popping language C > in mpicc -o conftest -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > conftest.o > Source: > #include "confdefs.h" > #include "conffix.h" > > int main() { > ; > return 0; > } > Rejecting linker flag -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 due > to nonzero status from link > Rejecting C linker flag -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 > due to > i686-apple-darwin8-gcc-4.0.1: unrecognized option '-rpath' > /usr/bin/ld: can't map file: /Users/bknaepen/Desktop/petsc-2.3.3-p6 ((os/kern) > invalid argument) > collect2: ld returned 1 exit status > > Rejected C linker flag -rpath > Trying C linker flag -Wl,-R, > sh: mpicc -c -o conftest.o -fPIC conftest.c > Executing: mpicc -c -o conftest.o -fPIC conftest.c > sh: > Pushing language C > Popping language C > sh: mpicc -o conftest -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > conftest.o > Executing: mpicc -o conftest -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > conftest.o > sh: > Possible ERROR while running linker: /usr/bin/ld: unknown flag: -R > collect2: ld returned 1 exit status > output: ret = 256 > error message = {/usr/bin/ld: unknown flag: -R > collect2: ld returned 1 exit status > } > Pushing language C > Popping language C > in mpicc -o conftest -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > conftest.o > Source: > #include "confdefs.h" > #include "conffix.h" > > int main() { > ; > return 0; > } > Rejecting linker flag -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 due > to nonzero status from link > Rejecting C linker flag -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > due to > /usr/bin/ld: unknown flag: -R > collect2: ld returned 1 exit status > > Rejected C linker flag -Wl,-R, > Popping language C > Pushing language FC > sh: mpicc -flags > Executing: mpicc -flags > sh: > Trying FC linker flag -Wl,-rpath, > sh: mpif90 -c -o conftest.o -I. -fPIC conftest.F > Executing: mpif90 -c -o conftest.o -I. -fPIC conftest.F > sh: > Pushing language C > Popping language C > sh: mpif90 -o conftest -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > -I. -fPIC conftest.o > Executing: mpif90 -o conftest > -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. -fPIC conftest.o > sh: > Possible ERROR while running linker: ipo: error #11019: OBJREAD Error: Could > not create mapping for /Users/bknaepen/Desktop/petsc-2.3.3-p6 > ifort: error #10014: problem during multi-file optimization compilation (code > 1) > output: ret = 256 > error message = {ipo: error #11019: OBJREAD Error: Could not create mapping > for /Users/bknaepen/Desktop/petsc-2.3.3-p6 > ifort: error #10014: problem during multi-file optimization compilation (code > 1) > } > Pushing language C > Popping language C > in mpif90 -o conftest -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. > -fPIC conftest.o > Source: > program main > > end > Rejecting linker flag -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > due to nonzero status from link > Valid FC linker flag -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > Rejected FC linker flag -Wl,-rpath, > Trying FC linker flag -R > sh: mpif90 -c -o conftest.o -I. -fPIC conftest.F > Executing: mpif90 -c -o conftest.o -I. -fPIC conftest.F > sh: > Pushing language C > Popping language C > sh: mpif90 -o conftest -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. -fPIC > conftest.o > Executing: mpif90 -o conftest -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. > -fPIC conftest.o > sh: > Possible ERROR while running linker: ifort: command line warning #10006: > ignoring unknown option '-R/Users/bknaepen/Desktop/petsc-2.3.3-p6' > output: error message = {ifort: command line warning #10006: ignoring unknown > option '-R/Users/bknaepen/Desktop/petsc-2.3.3-p6' > } > Pushing language C > Popping language C > in mpif90 -o conftest -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. -fPIC > conftest.o > Source: > program main > > end > Rejecting FC linker flag -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 due > to > ifort: command line warning #10006: ignoring unknown option > '-R/Users/bknaepen/Desktop/petsc-2.3.3-p6' > > Rejected FC linker flag -R > Trying FC linker flag -rpath > sh: mpif90 -c -o conftest.o -I. -fPIC conftest.F > Executing: mpif90 -c -o conftest.o -I. -fPIC conftest.F > sh: > Pushing language C > Popping language C > sh: mpif90 -o conftest -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. > -fPIC conftest.o > Executing: mpif90 -o conftest -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 > -I. -fPIC conftest.o > Runaway process exceeded time limit of 600.0s > Possible ERROR while running linker: Runaway process exceeded time limit of > 600.0s > output: Runaway processret = -1 > error message = {Runaway process exceeded time limit of 600.0s > } > Pushing language C > Popping language C > in mpif90 -o conftest -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. > -fPIC conftest.o > Source: > program main > > end > Rejecting linker flag -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 due > to nonzero status from link > Valid FC linker flag -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 > Rejected FC linker flag -rpath > Trying FC linker flag -Wl,-R, > sh: mpif90 -c -o conftest.o -I. -fPIC conftest.F > Executing: mpif90 -c -o conftest.o -I. -fPIC conftest.F > sh: > Pushing language C > Popping language C > sh: mpif90 -o conftest -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. > -fPIC conftest.o > Executing: mpif90 -o conftest -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > -I. -fPIC conftest.o > sh: > Possible ERROR while running linker: IPO link: Warning unknown option '-R'. > ipo: error #11019: OBJREAD Error: Could not create mapping for > /Users/bknaepen/Desktop/petsc-2.3.3-p6 > ifort: error #10014: problem during multi-file optimization compilation (code > 1) > output: ret = 256 > error message = {IPO link: Warning unknown option '-R'. > ipo: error #11019: OBJREAD Error: Could not create mapping for > /Users/bknaepen/Desktop/petsc-2.3.3-p6 > ifort: error #10014: problem during multi-file optimization compilation (code > 1) > } > Pushing language C > Popping language C > in mpif90 -o conftest -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. > -fPIC conftest.o > Source: > program main > > end > Rejecting linker flag -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 due > to nonzero status from link > Rejecting FC linker flag -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > due to > IPO link: Warning unknown option '-R'. > ipo: error #11019: OBJREAD Error: Could not create mapping for > /Users/bknaepen/Desktop/petsc-2.3.3-p6 > ifort: error #10014: problem during multi-file optimization compilation (code > 1) > > Rejected FC linker flag -Wl,-R, > Popping language FC > ================================================================================ > On 12 Oct 2007, at 23:26, Barry Smith wrote: > > > > > I have run with 10.0.0.20 on the Mac with no difficulties using the options > > below with petsc-2.3.3-6 and have no hanging. config/configure.py just takes > > a > > couple of minutes. > > > > [bsmith:~/Src/petsc-release-2.3.3] bsmith% uname -a > > Darwin bsmith.mcs.anl.gov 8.10.1 Darwin Kernel Version 8.10.1: Wed May 23 > > 16:33:00 PDT 2007; root:xnu-792.22.5~1/RELEASE_I386 i386 i386 > > > > > > Barry > > > > > > On Fri, 12 Oct 2007, Knaepen Bernard wrote: > > > > > Hello, > > > > > > is it normal that the ./config/configure.py stalls for a long time when > > > building petsc-2.3.3-p6 on MacOSX? The step at which it stays stuck is: > > > > > > TESTING: checkSharedLinkerPaths from > > > config.setCompilers(python/BuildSystem/config/setCompilers.py:1107) > > > > > > It remains there for around five minutes then continues the configure > > > process without problem. Except that it leaves a fortran process alive > > > that I have to manually kill later on. This process looks like: > > > > > > 17623 fortcom 60.8% 5:06.44 1 13 60 364K 12.1M 2.23M > > > 57.8M > > > 17614 ifortia32 0.0% 0:00.02 1 14 19 392K 1.96M 1.86M > > > 29.0M > > > 17613 ifortbin 0.0% 0:00.01 1 13 16 148K 992K 876K > > > 28.6M > > > > > > when I use the top command. > > > > > > I am configuring with the following options: > > > > > > /config/configure.py --with-cc=mpicc --with-fc=mpif90 --with-hypre=1 > > > --download-hypre=1 --with-debugging=0 COPTFLAGS=-O3 FOPTFLAGS=-O3 > > > --with-cxx=mpicxx --with-x=0 --with-sudo=sudo > > > > > > The system is MacOSX 10.4.9 (latest) running on a MacPro (or Macbook Pro) > > > and the fortran compiler is ifort 10.0.17. > > > > > > Regards, > > > Bernard. > > > > > > > > > > > > > From balay at mcs.anl.gov Sat Oct 13 19:49:23 2007 From: balay at mcs.anl.gov (Satish Balay) Date: Sat, 13 Oct 2007 19:49:23 -0500 (CDT) Subject: configure stalling on MacOSX In-Reply-To: References: <20071012072820.1D5D16AE@bonito.ulb.ac.be> Message-ID: Just a note: I've noticed similar hang with 'icc' on Dinesh's Mac. The thing that hangs is some component within the compiler -not compile wrapper. This is noticeable with graphical process listing utility from OSX' [I managed to kill it from this tool for configure to continue]. I didn't spend enough time debugging this. However I'm pretty sure [in case of icc] its a compiler bug - not any MPI wrapper [I was using --with-mpi=0 for this test] Will try to dig more when I get a chance. [perhaps sometime nextweek]. Satish On Sat, 13 Oct 2007, Barry Smith wrote: > > You get > > > Executing: mpif90 -o conftest -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. -fPIC conftest.o > > Runaway process exceeded time limit of 600.0s > > Possible ERROR while running linker: Runaway process exceeded time limit of 600.0s > > output: Runaway processret = -1 > > error message = {Runaway process exceeded time limit of 600.0s } > > Pushing language C > > Popping language C > > I get (not using mpif90, whose is it anyways MPICH1, MPICH2, Lam, OpenMPI?) > > sh: ifort -o conftest -rpath /Users/bsmith/Src/petsc-dev -fPIC conftest.o > Executing: ifort -o conftest -rpath /Users/bsmith/Src/petsc-dev -fPIC conftest.o > sh: > Possible ERROR while running linker: ifort: command line warning #10156: ignoring option '-r'; no argument required > ifort: error #10106: Fatal error in /opt/intel/fc/10.0.020/bin/fortcom, terminated by bus error > ifort: error #10014: problem during multi-file optimization compilation (code 1) > output: ret = 256 > > My guess is that your underlying ifort is crashing with a bus error just like > mine, but your mpif90 is not properly handling this and so doesn't end and return like it should. > > Can you try running the compile line that hangs directly from the shell with both ifort > directly and then your mpif90; what happens in each case? (Just make a tiny Fortran program in a file). > > Perhaps yet another reason for hating the absurd "MPI compiler wrappers"; like MPI is SO-so-so special > (because it needs to link against a couple of MPI libraries and maybe a couple of system libraries?) > that it cannot be be compiled and linked like EVERY OTHER F*&& Unix package ever written but needs > its own special compiler wrappers, make me sick. > > Barry > > > On Sat, 13 Oct 2007, Bernard Knaepen wrote: > > > Hi Matthew and Barry, > > > > I tried again after installing intel fortran 10.0.020 but I still get the > > hanging. I copy below the output from configure.log that exhibits the timeout. > > When I remove the test related to '-rpath ' then the scripts does not hang. No > > problems either if I disable fortran with --with-fc=0 > > > > uname -a gives: > > > > Darwin dolfin.local 8.10.1 Darwin Kernel Version 8.10.1: Wed May 23 16:33:00 > > PDT 2007; root:xnu-792.22.5~1/RELEASE_I386 i386 i386 > > > > Regards, > > Bernard. > > > > > > > > ================================================================================ > > TEST checkSharedLinkerPaths from > > config.setCompilers(/Users/bknaepen/Desktop/petsc-2.3.3-p6/python/BuildSystem/config/setCompilers.py:1107) > > TESTING: checkSharedLinkerPaths from > > config.setCompilers(python/BuildSystem/config/setCompilers.py:1107) > > Determine the shared linker path options > > - IRIX: -rpath > > - Linux, OSF: -Wl,-rpath, > > - Solaris: -R > > - FreeBSD: -Wl,-R, > > Pushing language C > > sh: mpicc -flags > > Executing: mpicc -flags > > sh: > > Trying C linker flag -Wl,-rpath, > > sh: mpicc -c -o conftest.o -fPIC conftest.c > > Executing: mpicc -c -o conftest.o -fPIC conftest.c > > sh: > > Pushing language C > > Popping language C > > sh: mpicc -o conftest -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > > conftest.o > > Executing: mpicc -o conftest > > -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > > conftest.o > > sh: > > Possible ERROR while running linker: /usr/bin/ld: unknown flag: -rpath > > collect2: ld returned 1 exit status > > output: ret = 256 > > error message = {/usr/bin/ld: unknown flag: -rpath > > collect2: ld returned 1 exit status > > } > > Pushing language C > > Popping language C > > in mpicc -o conftest -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > > conftest.o > > Source: > > #include "confdefs.h" > > #include "conffix.h" > > > > int main() { > > ; > > return 0; > > } > > Rejecting linker flag -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > due to nonzero status from link > > Rejecting C linker flag > > -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 due to > > /usr/bin/ld: unknown flag: -rpath > > collect2: ld returned 1 exit status > > > > Rejected C linker flag -Wl,-rpath, > > Trying C linker flag -R > > sh: mpicc -c -o conftest.o -fPIC conftest.c > > Executing: mpicc -c -o conftest.o -fPIC conftest.c > > sh: > > Pushing language C > > Popping language C > > sh: mpicc -o conftest -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > > conftest.o > > Executing: mpicc -o conftest -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > > conftest.o > > sh: > > Possible ERROR while running linker: i686-apple-darwin8-gcc-4.0.1: > > unrecognized option '-R/Users/bknaepen/Desktop/petsc-2.3.3-p6' > > output: error message = {i686-apple-darwin8-gcc-4.0.1: unrecognized option > > '-R/Users/bknaepen/Desktop/petsc-2.3.3-p6' > > } > > Pushing language C > > Popping language C > > in mpicc -o conftest -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > > conftest.o > > Source: > > #include "confdefs.h" > > #include "conffix.h" > > > > int main() { > > ; > > return 0; > > } > > Rejecting C linker flag -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 due to > > i686-apple-darwin8-gcc-4.0.1: unrecognized option > > '-R/Users/bknaepen/Desktop/petsc-2.3.3-p6' > > > > Rejected C linker flag -R > > Trying C linker flag -rpath > > sh: mpicc -c -o conftest.o -fPIC conftest.c > > Executing: mpicc -c -o conftest.o -fPIC conftest.c > > sh: > > Pushing language C > > Popping language C > > sh: mpicc -o conftest -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > > conftest.o > > Executing: mpicc -o conftest -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > > conftest.o > > sh: > > Possible ERROR while running linker: i686-apple-darwin8-gcc-4.0.1: > > unrecognized option '-rpath' > > /usr/bin/ld: can't map file: /Users/bknaepen/Desktop/petsc-2.3.3-p6 ((os/kern) > > invalid argument) > > collect2: ld returned 1 exit status > > output: ret = 256 > > error message = {i686-apple-darwin8-gcc-4.0.1: unrecognized option '-rpath' > > /usr/bin/ld: can't map file: /Users/bknaepen/Desktop/petsc-2.3.3-p6 ((os/kern) > > invalid argument) > > collect2: ld returned 1 exit status > > } > > Pushing language C > > Popping language C > > in mpicc -o conftest -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > > conftest.o > > Source: > > #include "confdefs.h" > > #include "conffix.h" > > > > int main() { > > ; > > return 0; > > } > > Rejecting linker flag -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 due > > to nonzero status from link > > Rejecting C linker flag -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 > > due to > > i686-apple-darwin8-gcc-4.0.1: unrecognized option '-rpath' > > /usr/bin/ld: can't map file: /Users/bknaepen/Desktop/petsc-2.3.3-p6 ((os/kern) > > invalid argument) > > collect2: ld returned 1 exit status > > > > Rejected C linker flag -rpath > > Trying C linker flag -Wl,-R, > > sh: mpicc -c -o conftest.o -fPIC conftest.c > > Executing: mpicc -c -o conftest.o -fPIC conftest.c > > sh: > > Pushing language C > > Popping language C > > sh: mpicc -o conftest -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > > conftest.o > > Executing: mpicc -o conftest -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > > conftest.o > > sh: > > Possible ERROR while running linker: /usr/bin/ld: unknown flag: -R > > collect2: ld returned 1 exit status > > output: ret = 256 > > error message = {/usr/bin/ld: unknown flag: -R > > collect2: ld returned 1 exit status > > } > > Pushing language C > > Popping language C > > in mpicc -o conftest -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > > conftest.o > > Source: > > #include "confdefs.h" > > #include "conffix.h" > > > > int main() { > > ; > > return 0; > > } > > Rejecting linker flag -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 due > > to nonzero status from link > > Rejecting C linker flag -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > due to > > /usr/bin/ld: unknown flag: -R > > collect2: ld returned 1 exit status > > > > Rejected C linker flag -Wl,-R, > > Popping language C > > Pushing language FC > > sh: mpicc -flags > > Executing: mpicc -flags > > sh: > > Trying FC linker flag -Wl,-rpath, > > sh: mpif90 -c -o conftest.o -I. -fPIC conftest.F > > Executing: mpif90 -c -o conftest.o -I. -fPIC conftest.F > > sh: > > Pushing language C > > Popping language C > > sh: mpif90 -o conftest -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > -I. -fPIC conftest.o > > Executing: mpif90 -o conftest > > -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. -fPIC conftest.o > > sh: > > Possible ERROR while running linker: ipo: error #11019: OBJREAD Error: Could > > not create mapping for /Users/bknaepen/Desktop/petsc-2.3.3-p6 > > ifort: error #10014: problem during multi-file optimization compilation (code > > 1) > > output: ret = 256 > > error message = {ipo: error #11019: OBJREAD Error: Could not create mapping > > for /Users/bknaepen/Desktop/petsc-2.3.3-p6 > > ifort: error #10014: problem during multi-file optimization compilation (code > > 1) > > } > > Pushing language C > > Popping language C > > in mpif90 -o conftest -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. > > -fPIC conftest.o > > Source: > > program main > > > > end > > Rejecting linker flag -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > due to nonzero status from link > > Valid FC linker flag -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > Rejected FC linker flag -Wl,-rpath, > > Trying FC linker flag -R > > sh: mpif90 -c -o conftest.o -I. -fPIC conftest.F > > Executing: mpif90 -c -o conftest.o -I. -fPIC conftest.F > > sh: > > Pushing language C > > Popping language C > > sh: mpif90 -o conftest -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. -fPIC > > conftest.o > > Executing: mpif90 -o conftest -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. > > -fPIC conftest.o > > sh: > > Possible ERROR while running linker: ifort: command line warning #10006: > > ignoring unknown option '-R/Users/bknaepen/Desktop/petsc-2.3.3-p6' > > output: error message = {ifort: command line warning #10006: ignoring unknown > > option '-R/Users/bknaepen/Desktop/petsc-2.3.3-p6' > > } > > Pushing language C > > Popping language C > > in mpif90 -o conftest -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. -fPIC > > conftest.o > > Source: > > program main > > > > end > > Rejecting FC linker flag -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 due > > to > > ifort: command line warning #10006: ignoring unknown option > > '-R/Users/bknaepen/Desktop/petsc-2.3.3-p6' > > > > Rejected FC linker flag -R > > Trying FC linker flag -rpath > > sh: mpif90 -c -o conftest.o -I. -fPIC conftest.F > > Executing: mpif90 -c -o conftest.o -I. -fPIC conftest.F > > sh: > > Pushing language C > > Popping language C > > sh: mpif90 -o conftest -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. > > -fPIC conftest.o > > Executing: mpif90 -o conftest -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 > > -I. -fPIC conftest.o > > Runaway process exceeded time limit of 600.0s > > Possible ERROR while running linker: Runaway process exceeded time limit of > > 600.0s > > output: Runaway processret = -1 > > error message = {Runaway process exceeded time limit of 600.0s > > } > > Pushing language C > > Popping language C > > in mpif90 -o conftest -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. > > -fPIC conftest.o > > Source: > > program main > > > > end > > Rejecting linker flag -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 due > > to nonzero status from link > > Valid FC linker flag -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 > > Rejected FC linker flag -rpath > > Trying FC linker flag -Wl,-R, > > sh: mpif90 -c -o conftest.o -I. -fPIC conftest.F > > Executing: mpif90 -c -o conftest.o -I. -fPIC conftest.F > > sh: > > Pushing language C > > Popping language C > > sh: mpif90 -o conftest -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. > > -fPIC conftest.o > > Executing: mpif90 -o conftest -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > -I. -fPIC conftest.o > > sh: > > Possible ERROR while running linker: IPO link: Warning unknown option '-R'. > > ipo: error #11019: OBJREAD Error: Could not create mapping for > > /Users/bknaepen/Desktop/petsc-2.3.3-p6 > > ifort: error #10014: problem during multi-file optimization compilation (code > > 1) > > output: ret = 256 > > error message = {IPO link: Warning unknown option '-R'. > > ipo: error #11019: OBJREAD Error: Could not create mapping for > > /Users/bknaepen/Desktop/petsc-2.3.3-p6 > > ifort: error #10014: problem during multi-file optimization compilation (code > > 1) > > } > > Pushing language C > > Popping language C > > in mpif90 -o conftest -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. > > -fPIC conftest.o > > Source: > > program main > > > > end > > Rejecting linker flag -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 due > > to nonzero status from link > > Rejecting FC linker flag -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > due to > > IPO link: Warning unknown option '-R'. > > ipo: error #11019: OBJREAD Error: Could not create mapping for > > /Users/bknaepen/Desktop/petsc-2.3.3-p6 > > ifort: error #10014: problem during multi-file optimization compilation (code > > 1) > > > > Rejected FC linker flag -Wl,-R, > > Popping language FC > > ================================================================================ > > On 12 Oct 2007, at 23:26, Barry Smith wrote: > > > > > > > > I have run with 10.0.0.20 on the Mac with no difficulties using the options > > > below with petsc-2.3.3-6 and have no hanging. config/configure.py just takes > > > a > > > couple of minutes. > > > > > > [bsmith:~/Src/petsc-release-2.3.3] bsmith% uname -a > > > Darwin bsmith.mcs.anl.gov 8.10.1 Darwin Kernel Version 8.10.1: Wed May 23 > > > 16:33:00 PDT 2007; root:xnu-792.22.5~1/RELEASE_I386 i386 i386 > > > > > > > > > Barry > > > > > > > > > On Fri, 12 Oct 2007, Knaepen Bernard wrote: > > > > > > > Hello, > > > > > > > > is it normal that the ./config/configure.py stalls for a long time when > > > > building petsc-2.3.3-p6 on MacOSX? The step at which it stays stuck is: > > > > > > > > TESTING: checkSharedLinkerPaths from > > > > config.setCompilers(python/BuildSystem/config/setCompilers.py:1107) > > > > > > > > It remains there for around five minutes then continues the configure > > > > process without problem. Except that it leaves a fortran process alive > > > > that I have to manually kill later on. This process looks like: > > > > > > > > 17623 fortcom 60.8% 5:06.44 1 13 60 364K 12.1M 2.23M > > > > 57.8M > > > > 17614 ifortia32 0.0% 0:00.02 1 14 19 392K 1.96M 1.86M > > > > 29.0M > > > > 17613 ifortbin 0.0% 0:00.01 1 13 16 148K 992K 876K > > > > 28.6M > > > > > > > > when I use the top command. > > > > > > > > I am configuring with the following options: > > > > > > > > /config/configure.py --with-cc=mpicc --with-fc=mpif90 --with-hypre=1 > > > > --download-hypre=1 --with-debugging=0 COPTFLAGS=-O3 FOPTFLAGS=-O3 > > > > --with-cxx=mpicxx --with-x=0 --with-sudo=sudo > > > > > > > > The system is MacOSX 10.4.9 (latest) running on a MacPro (or Macbook Pro) > > > > and the fortran compiler is ifort 10.0.17. > > > > > > > > Regards, > > > > Bernard. > > > > > > > > > > > > > > > > > > > > From bsmith at mcs.anl.gov Sun Oct 14 10:24:25 2007 From: bsmith at mcs.anl.gov (Barry Smith) Date: Sun, 14 Oct 2007 10:24:25 -0500 (CDT) Subject: configure stalling on MacOSX In-Reply-To: References: <20071012072820.1D5D16AE@bonito.ulb.ac.be> Message-ID: Ok. It never hangs for me, for several different versions of icc/ifort. Must be a hardware problem :-) Barry On Sat, 13 Oct 2007, Satish Balay wrote: > Just a note: I've noticed similar hang with 'icc' on Dinesh's Mac. > The thing that hangs is some component within the compiler -not > compile wrapper. This is noticeable with graphical process listing > utility from OSX' [I managed to kill it from this tool for configure > to continue]. > > I didn't spend enough time debugging this. However I'm pretty sure [in > case of icc] its a compiler bug - not any MPI wrapper [I was using > --with-mpi=0 for this test] > > Will try to dig more when I get a chance. [perhaps sometime nextweek]. > > Satish > > On Sat, 13 Oct 2007, Barry Smith wrote: > > > > > You get > > > > > Executing: mpif90 -o conftest -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. -fPIC conftest.o > > > Runaway process exceeded time limit of 600.0s > > > Possible ERROR while running linker: Runaway process exceeded time limit of 600.0s > > > output: Runaway processret = -1 > > > error message = {Runaway process exceeded time limit of 600.0s } > > > Pushing language C > > > Popping language C > > > > I get (not using mpif90, whose is it anyways MPICH1, MPICH2, Lam, OpenMPI?) > > > > sh: ifort -o conftest -rpath /Users/bsmith/Src/petsc-dev -fPIC conftest.o > > Executing: ifort -o conftest -rpath /Users/bsmith/Src/petsc-dev -fPIC conftest.o > > sh: > > Possible ERROR while running linker: ifort: command line warning #10156: ignoring option '-r'; no argument required > > ifort: error #10106: Fatal error in /opt/intel/fc/10.0.020/bin/fortcom, terminated by bus error > > ifort: error #10014: problem during multi-file optimization compilation (code 1) > > output: ret = 256 > > > > My guess is that your underlying ifort is crashing with a bus error just like > > mine, but your mpif90 is not properly handling this and so doesn't end and return like it should. > > > > Can you try running the compile line that hangs directly from the shell with both ifort > > directly and then your mpif90; what happens in each case? (Just make a tiny Fortran program in a file). > > > > Perhaps yet another reason for hating the absurd "MPI compiler wrappers"; like MPI is SO-so-so special > > (because it needs to link against a couple of MPI libraries and maybe a couple of system libraries?) > > that it cannot be be compiled and linked like EVERY OTHER F*&& Unix package ever written but needs > > its own special compiler wrappers, make me sick. > > > > Barry > > > > > > On Sat, 13 Oct 2007, Bernard Knaepen wrote: > > > > > Hi Matthew and Barry, > > > > > > I tried again after installing intel fortran 10.0.020 but I still get the > > > hanging. I copy below the output from configure.log that exhibits the timeout. > > > When I remove the test related to '-rpath ' then the scripts does not hang. No > > > problems either if I disable fortran with --with-fc=0 > > > > > > uname -a gives: > > > > > > Darwin dolfin.local 8.10.1 Darwin Kernel Version 8.10.1: Wed May 23 16:33:00 > > > PDT 2007; root:xnu-792.22.5~1/RELEASE_I386 i386 i386 > > > > > > Regards, > > > Bernard. > > > > > > > > > > > > ================================================================================ > > > TEST checkSharedLinkerPaths from > > > config.setCompilers(/Users/bknaepen/Desktop/petsc-2.3.3-p6/python/BuildSystem/config/setCompilers.py:1107) > > > TESTING: checkSharedLinkerPaths from > > > config.setCompilers(python/BuildSystem/config/setCompilers.py:1107) > > > Determine the shared linker path options > > > - IRIX: -rpath > > > - Linux, OSF: -Wl,-rpath, > > > - Solaris: -R > > > - FreeBSD: -Wl,-R, > > > Pushing language C > > > sh: mpicc -flags > > > Executing: mpicc -flags > > > sh: > > > Trying C linker flag -Wl,-rpath, > > > sh: mpicc -c -o conftest.o -fPIC conftest.c > > > Executing: mpicc -c -o conftest.o -fPIC conftest.c > > > sh: > > > Pushing language C > > > Popping language C > > > sh: mpicc -o conftest -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > > > conftest.o > > > Executing: mpicc -o conftest > > > -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > > > conftest.o > > > sh: > > > Possible ERROR while running linker: /usr/bin/ld: unknown flag: -rpath > > > collect2: ld returned 1 exit status > > > output: ret = 256 > > > error message = {/usr/bin/ld: unknown flag: -rpath > > > collect2: ld returned 1 exit status > > > } > > > Pushing language C > > > Popping language C > > > in mpicc -o conftest -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > > > conftest.o > > > Source: > > > #include "confdefs.h" > > > #include "conffix.h" > > > > > > int main() { > > > ; > > > return 0; > > > } > > > Rejecting linker flag -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > > due to nonzero status from link > > > Rejecting C linker flag > > > -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 due to > > > /usr/bin/ld: unknown flag: -rpath > > > collect2: ld returned 1 exit status > > > > > > Rejected C linker flag -Wl,-rpath, > > > Trying C linker flag -R > > > sh: mpicc -c -o conftest.o -fPIC conftest.c > > > Executing: mpicc -c -o conftest.o -fPIC conftest.c > > > sh: > > > Pushing language C > > > Popping language C > > > sh: mpicc -o conftest -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > > > conftest.o > > > Executing: mpicc -o conftest -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > > > conftest.o > > > sh: > > > Possible ERROR while running linker: i686-apple-darwin8-gcc-4.0.1: > > > unrecognized option '-R/Users/bknaepen/Desktop/petsc-2.3.3-p6' > > > output: error message = {i686-apple-darwin8-gcc-4.0.1: unrecognized option > > > '-R/Users/bknaepen/Desktop/petsc-2.3.3-p6' > > > } > > > Pushing language C > > > Popping language C > > > in mpicc -o conftest -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > > > conftest.o > > > Source: > > > #include "confdefs.h" > > > #include "conffix.h" > > > > > > int main() { > > > ; > > > return 0; > > > } > > > Rejecting C linker flag -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 due to > > > i686-apple-darwin8-gcc-4.0.1: unrecognized option > > > '-R/Users/bknaepen/Desktop/petsc-2.3.3-p6' > > > > > > Rejected C linker flag -R > > > Trying C linker flag -rpath > > > sh: mpicc -c -o conftest.o -fPIC conftest.c > > > Executing: mpicc -c -o conftest.o -fPIC conftest.c > > > sh: > > > Pushing language C > > > Popping language C > > > sh: mpicc -o conftest -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 > > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > > > conftest.o > > > Executing: mpicc -o conftest -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 > > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > > > conftest.o > > > sh: > > > Possible ERROR while running linker: i686-apple-darwin8-gcc-4.0.1: > > > unrecognized option '-rpath' > > > /usr/bin/ld: can't map file: /Users/bknaepen/Desktop/petsc-2.3.3-p6 ((os/kern) > > > invalid argument) > > > collect2: ld returned 1 exit status > > > output: ret = 256 > > > error message = {i686-apple-darwin8-gcc-4.0.1: unrecognized option '-rpath' > > > /usr/bin/ld: can't map file: /Users/bknaepen/Desktop/petsc-2.3.3-p6 ((os/kern) > > > invalid argument) > > > collect2: ld returned 1 exit status > > > } > > > Pushing language C > > > Popping language C > > > in mpicc -o conftest -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 > > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > > > conftest.o > > > Source: > > > #include "confdefs.h" > > > #include "conffix.h" > > > > > > int main() { > > > ; > > > return 0; > > > } > > > Rejecting linker flag -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 due > > > to nonzero status from link > > > Rejecting C linker flag -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 > > > due to > > > i686-apple-darwin8-gcc-4.0.1: unrecognized option '-rpath' > > > /usr/bin/ld: can't map file: /Users/bknaepen/Desktop/petsc-2.3.3-p6 ((os/kern) > > > invalid argument) > > > collect2: ld returned 1 exit status > > > > > > Rejected C linker flag -rpath > > > Trying C linker flag -Wl,-R, > > > sh: mpicc -c -o conftest.o -fPIC conftest.c > > > Executing: mpicc -c -o conftest.o -fPIC conftest.c > > > sh: > > > Pushing language C > > > Popping language C > > > sh: mpicc -o conftest -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > > > conftest.o > > > Executing: mpicc -o conftest -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > > > conftest.o > > > sh: > > > Possible ERROR while running linker: /usr/bin/ld: unknown flag: -R > > > collect2: ld returned 1 exit status > > > output: ret = 256 > > > error message = {/usr/bin/ld: unknown flag: -R > > > collect2: ld returned 1 exit status > > > } > > > Pushing language C > > > Popping language C > > > in mpicc -o conftest -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress > > > -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress -fPIC > > > conftest.o > > > Source: > > > #include "confdefs.h" > > > #include "conffix.h" > > > > > > int main() { > > > ; > > > return 0; > > > } > > > Rejecting linker flag -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 due > > > to nonzero status from link > > > Rejecting C linker flag -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > > due to > > > /usr/bin/ld: unknown flag: -R > > > collect2: ld returned 1 exit status > > > > > > Rejected C linker flag -Wl,-R, > > > Popping language C > > > Pushing language FC > > > sh: mpicc -flags > > > Executing: mpicc -flags > > > sh: > > > Trying FC linker flag -Wl,-rpath, > > > sh: mpif90 -c -o conftest.o -I. -fPIC conftest.F > > > Executing: mpif90 -c -o conftest.o -I. -fPIC conftest.F > > > sh: > > > Pushing language C > > > Popping language C > > > sh: mpif90 -o conftest -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > > -I. -fPIC conftest.o > > > Executing: mpif90 -o conftest > > > -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. -fPIC conftest.o > > > sh: > > > Possible ERROR while running linker: ipo: error #11019: OBJREAD Error: Could > > > not create mapping for /Users/bknaepen/Desktop/petsc-2.3.3-p6 > > > ifort: error #10014: problem during multi-file optimization compilation (code > > > 1) > > > output: ret = 256 > > > error message = {ipo: error #11019: OBJREAD Error: Could not create mapping > > > for /Users/bknaepen/Desktop/petsc-2.3.3-p6 > > > ifort: error #10014: problem during multi-file optimization compilation (code > > > 1) > > > } > > > Pushing language C > > > Popping language C > > > in mpif90 -o conftest -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. > > > -fPIC conftest.o > > > Source: > > > program main > > > > > > end > > > Rejecting linker flag -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > > due to nonzero status from link > > > Valid FC linker flag -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > > Rejected FC linker flag -Wl,-rpath, > > > Trying FC linker flag -R > > > sh: mpif90 -c -o conftest.o -I. -fPIC conftest.F > > > Executing: mpif90 -c -o conftest.o -I. -fPIC conftest.F > > > sh: > > > Pushing language C > > > Popping language C > > > sh: mpif90 -o conftest -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. -fPIC > > > conftest.o > > > Executing: mpif90 -o conftest -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. > > > -fPIC conftest.o > > > sh: > > > Possible ERROR while running linker: ifort: command line warning #10006: > > > ignoring unknown option '-R/Users/bknaepen/Desktop/petsc-2.3.3-p6' > > > output: error message = {ifort: command line warning #10006: ignoring unknown > > > option '-R/Users/bknaepen/Desktop/petsc-2.3.3-p6' > > > } > > > Pushing language C > > > Popping language C > > > in mpif90 -o conftest -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. -fPIC > > > conftest.o > > > Source: > > > program main > > > > > > end > > > Rejecting FC linker flag -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 due > > > to > > > ifort: command line warning #10006: ignoring unknown option > > > '-R/Users/bknaepen/Desktop/petsc-2.3.3-p6' > > > > > > Rejected FC linker flag -R > > > Trying FC linker flag -rpath > > > sh: mpif90 -c -o conftest.o -I. -fPIC conftest.F > > > Executing: mpif90 -c -o conftest.o -I. -fPIC conftest.F > > > sh: > > > Pushing language C > > > Popping language C > > > sh: mpif90 -o conftest -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. > > > -fPIC conftest.o > > > Executing: mpif90 -o conftest -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 > > > -I. -fPIC conftest.o > > > Runaway process exceeded time limit of 600.0s > > > Possible ERROR while running linker: Runaway process exceeded time limit of > > > 600.0s > > > output: Runaway processret = -1 > > > error message = {Runaway process exceeded time limit of 600.0s > > > } > > > Pushing language C > > > Popping language C > > > in mpif90 -o conftest -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. > > > -fPIC conftest.o > > > Source: > > > program main > > > > > > end > > > Rejecting linker flag -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 due > > > to nonzero status from link > > > Valid FC linker flag -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 > > > Rejected FC linker flag -rpath > > > Trying FC linker flag -Wl,-R, > > > sh: mpif90 -c -o conftest.o -I. -fPIC conftest.F > > > Executing: mpif90 -c -o conftest.o -I. -fPIC conftest.F > > > sh: > > > Pushing language C > > > Popping language C > > > sh: mpif90 -o conftest -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. > > > -fPIC conftest.o > > > Executing: mpif90 -o conftest -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > > -I. -fPIC conftest.o > > > sh: > > > Possible ERROR while running linker: IPO link: Warning unknown option '-R'. > > > ipo: error #11019: OBJREAD Error: Could not create mapping for > > > /Users/bknaepen/Desktop/petsc-2.3.3-p6 > > > ifort: error #10014: problem during multi-file optimization compilation (code > > > 1) > > > output: ret = 256 > > > error message = {IPO link: Warning unknown option '-R'. > > > ipo: error #11019: OBJREAD Error: Could not create mapping for > > > /Users/bknaepen/Desktop/petsc-2.3.3-p6 > > > ifort: error #10014: problem during multi-file optimization compilation (code > > > 1) > > > } > > > Pushing language C > > > Popping language C > > > in mpif90 -o conftest -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. > > > -fPIC conftest.o > > > Source: > > > program main > > > > > > end > > > Rejecting linker flag -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 due > > > to nonzero status from link > > > Rejecting FC linker flag -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 > > > due to > > > IPO link: Warning unknown option '-R'. > > > ipo: error #11019: OBJREAD Error: Could not create mapping for > > > /Users/bknaepen/Desktop/petsc-2.3.3-p6 > > > ifort: error #10014: problem during multi-file optimization compilation (code > > > 1) > > > > > > Rejected FC linker flag -Wl,-R, > > > Popping language FC > > > ================================================================================ > > > On 12 Oct 2007, at 23:26, Barry Smith wrote: > > > > > > > > > > > I have run with 10.0.0.20 on the Mac with no difficulties using the options > > > > below with petsc-2.3.3-6 and have no hanging. config/configure.py just takes > > > > a > > > > couple of minutes. > > > > > > > > [bsmith:~/Src/petsc-release-2.3.3] bsmith% uname -a > > > > Darwin bsmith.mcs.anl.gov 8.10.1 Darwin Kernel Version 8.10.1: Wed May 23 > > > > 16:33:00 PDT 2007; root:xnu-792.22.5~1/RELEASE_I386 i386 i386 > > > > > > > > > > > > Barry > > > > > > > > > > > > On Fri, 12 Oct 2007, Knaepen Bernard wrote: > > > > > > > > > Hello, > > > > > > > > > > is it normal that the ./config/configure.py stalls for a long time when > > > > > building petsc-2.3.3-p6 on MacOSX? The step at which it stays stuck is: > > > > > > > > > > TESTING: checkSharedLinkerPaths from > > > > > config.setCompilers(python/BuildSystem/config/setCompilers.py:1107) > > > > > > > > > > It remains there for around five minutes then continues the configure > > > > > process without problem. Except that it leaves a fortran process alive > > > > > that I have to manually kill later on. This process looks like: > > > > > > > > > > 17623 fortcom 60.8% 5:06.44 1 13 60 364K 12.1M 2.23M > > > > > 57.8M > > > > > 17614 ifortia32 0.0% 0:00.02 1 14 19 392K 1.96M 1.86M > > > > > 29.0M > > > > > 17613 ifortbin 0.0% 0:00.01 1 13 16 148K 992K 876K > > > > > 28.6M > > > > > > > > > > when I use the top command. > > > > > > > > > > I am configuring with the following options: > > > > > > > > > > /config/configure.py --with-cc=mpicc --with-fc=mpif90 --with-hypre=1 > > > > > --download-hypre=1 --with-debugging=0 COPTFLAGS=-O3 FOPTFLAGS=-O3 > > > > > --with-cxx=mpicxx --with-x=0 --with-sudo=sudo > > > > > > > > > > The system is MacOSX 10.4.9 (latest) running on a MacPro (or Macbook Pro) > > > > > and the fortran compiler is ifort 10.0.17. > > > > > > > > > > Regards, > > > > > Bernard. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > From bknaepen at ulb.ac.be Sun Oct 14 12:56:53 2007 From: bknaepen at ulb.ac.be (Bernard Knaepen) Date: Sun, 14 Oct 2007 19:56:53 +0200 Subject: configure stalling on MacOSX In-Reply-To: References: <20071012072820.1D5D16AE@bonito.ulb.ac.be> Message-ID: <4E21B847-5F56-49E8-B3F9-EE685C5B3FB6@ulb.ac.be> Hi, If I try to compile with ifort a (helloworld) program directly from the command line with the rpath option I get: dolfin:~/Desktop bknaepen$ ifort -o test -rpath /Users/bknaepen/ Desktop/petsc-2.3.3-p6 -I. -fPIC test.o ifort: command line warning #10156: ignoring option '-r'; no argument required ifort: error #10106: Fatal error in /opt/intel/fc/10.0.020/bin/ fortcom, terminated by bus error ifort: error #10014: problem during multi-file optimization compilation (code 1) Same result with mpif90 (my version of mpi is mpich2-106): dolfin:~/Desktop bknaepen$ mpif90 -o test -rpath /Users/bknaepen/ Desktop/petsc-2.3.3-p6 -I. -fPIC test.o ifort: command line warning #10156: ignoring option '-r'; no argument required ifort: error #10106: Fatal error in /opt/intel/fc/10.0.020/bin/ fortcom, terminated by bus error ifort: error #10014: problem during multi-file optimization compilation (code 1) In other words, the compiler complains and exits immediately without getting stuck. Regards, Bernard. On 13 Oct 2007, at 22:13, Barry Smith wrote: > > You get > >> Executing: mpif90 -o conftest -rpath /Users/bknaepen/Desktop/ >> petsc-2.3.3-p6 -I. -fPIC conftest.o >> Runaway process exceeded time limit of 600.0s >> Possible ERROR while running linker: Runaway process exceeded time >> limit of 600.0s >> output: Runaway processret = -1 >> error message = {Runaway process exceeded time limit of 600.0s } >> Pushing language C >> Popping language C > > I get (not using mpif90, whose is it anyways MPICH1, MPICH2, Lam, > OpenMPI?) > > sh: ifort -o conftest -rpath /Users/bsmith/Src/petsc-dev -fPIC > conftest.o > Executing: ifort -o conftest -rpath /Users/bsmith/Src/petsc-dev - > fPIC conftest.o > sh: > Possible ERROR while running linker: ifort: command line warning > #10156: ignoring option '-r'; no argument required > ifort: error #10106: Fatal error in /opt/intel/fc/10.0.020/bin/ > fortcom, terminated by bus error > ifort: error #10014: problem during multi-file optimization > compilation (code 1) > output: ret = 256 > > My guess is that your underlying ifort is crashing with a bus > error just like > mine, but your mpif90 is not properly handling this and so doesn't > end and return like it should. > > Can you try running the compile line that hangs directly from the > shell with both ifort > directly and then your mpif90; what happens in each case? (Just > make a tiny Fortran program in a file). > > Perhaps yet another reason for hating the absurd "MPI compiler > wrappers"; like MPI is SO-so-so special > (because it needs to link against a couple of MPI libraries and > maybe a couple of system libraries?) > that it cannot be be compiled and linked like EVERY OTHER F*&& Unix > package ever written but needs > its own special compiler wrappers, make me sick. > > Barry > > > On Sat, 13 Oct 2007, Bernard Knaepen wrote: > >> Hi Matthew and Barry, >> >> I tried again after installing intel fortran 10.0.020 but I still >> get the >> hanging. I copy below the output from configure.log that exhibits >> the timeout. >> When I remove the test related to '-rpath ' then the scripts does >> not hang. No >> problems either if I disable fortran with --with-fc=0 >> >> uname -a gives: >> >> Darwin dolfin.local 8.10.1 Darwin Kernel Version 8.10.1: Wed May >> 23 16:33:00 >> PDT 2007; root:xnu-792.22.5~1/RELEASE_I386 i386 i386 >> >> Regards, >> Bernard. >> >> >> >> ===================================================================== >> =========== >> TEST checkSharedLinkerPaths from >> config.setCompilers(/Users/bknaepen/Desktop/petsc-2.3.3-p6/python/ >> BuildSystem/config/setCompilers.py:1107) >> TESTING: checkSharedLinkerPaths from >> config.setCompilers(python/BuildSystem/config/setCompilers.py:1107) >> Determine the shared linker path options >> - IRIX: -rpath >> - Linux, OSF: -Wl,-rpath, >> - Solaris: -R >> - FreeBSD: -Wl,-R, >> Pushing language C >> sh: mpicc -flags >> Executing: mpicc -flags >> sh: >> Trying C linker flag -Wl,-rpath, >> sh: mpicc -c -o conftest.o -fPIC conftest.c >> Executing: mpicc -c -o conftest.o -fPIC conftest.c >> sh: >> Pushing language C >> Popping language C >> sh: mpicc -o conftest -Wl,-rpath,/Users/bknaepen/Desktop/ >> petsc-2.3.3-p6 >> -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress >> -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress >> -fPIC >> conftest.o >> Executing: mpicc -o conftest >> -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 >> -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress >> -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress >> -fPIC >> conftest.o >> sh: >> Possible ERROR while running linker: /usr/bin/ld: unknown flag: - >> rpath >> collect2: ld returned 1 exit status >> output: ret = 256 >> error message = {/usr/bin/ld: unknown flag: -rpath >> collect2: ld returned 1 exit status >> } >> Pushing language C >> Popping language C >> in mpicc -o conftest -Wl,-rpath,/Users/bknaepen/Desktop/ >> petsc-2.3.3-p6 >> -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress >> -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress >> -fPIC >> conftest.o >> Source: >> #include "confdefs.h" >> #include "conffix.h" >> >> int main() { >> ; >> return 0; >> } >> Rejecting linker flag -Wl,-rpath,/Users/bknaepen/Desktop/ >> petsc-2.3.3-p6 >> due to nonzero status from link >> Rejecting C linker flag >> -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 due to >> /usr/bin/ld: unknown flag: -rpath >> collect2: ld returned 1 exit status >> >> Rejected C linker flag -Wl,-rpath, >> Trying C linker flag -R >> sh: mpicc -c -o conftest.o -fPIC conftest.c >> Executing: mpicc -c -o conftest.o -fPIC conftest.c >> sh: >> Pushing language C >> Popping language C >> sh: mpicc -o conftest -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 >> -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress >> -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress >> -fPIC >> conftest.o >> Executing: mpicc -o conftest -R/Users/bknaepen/Desktop/ >> petsc-2.3.3-p6 >> -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress >> -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress >> -fPIC >> conftest.o >> sh: >> Possible ERROR while running linker: i686-apple-darwin8-gcc-4.0.1: >> unrecognized option '-R/Users/bknaepen/Desktop/petsc-2.3.3-p6' >> output: error message = {i686-apple-darwin8-gcc-4.0.1: >> unrecognized option >> '-R/Users/bknaepen/Desktop/petsc-2.3.3-p6' >> } >> Pushing language C >> Popping language C >> in mpicc -o conftest -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 >> -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress >> -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress >> -fPIC >> conftest.o >> Source: >> #include "confdefs.h" >> #include "conffix.h" >> >> int main() { >> ; >> return 0; >> } >> Rejecting C linker flag -R/Users/bknaepen/Desktop/ >> petsc-2.3.3-p6 due to >> i686-apple-darwin8-gcc-4.0.1: unrecognized option >> '-R/Users/bknaepen/Desktop/petsc-2.3.3-p6' >> >> Rejected C linker flag -R >> Trying C linker flag -rpath >> sh: mpicc -c -o conftest.o -fPIC conftest.c >> Executing: mpicc -c -o conftest.o -fPIC conftest.c >> sh: >> Pushing language C >> Popping language C >> sh: mpicc -o conftest -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 >> -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress >> -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress >> -fPIC >> conftest.o >> Executing: mpicc -o conftest -rpath /Users/bknaepen/Desktop/ >> petsc-2.3.3-p6 >> -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress >> -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress >> -fPIC >> conftest.o >> sh: >> Possible ERROR while running linker: i686-apple-darwin8-gcc-4.0.1: >> unrecognized option '-rpath' >> /usr/bin/ld: can't map file: /Users/bknaepen/Desktop/petsc-2.3.3- >> p6 ((os/kern) >> invalid argument) >> collect2: ld returned 1 exit status >> output: ret = 256 >> error message = {i686-apple-darwin8-gcc-4.0.1: unrecognized option >> '-rpath' >> /usr/bin/ld: can't map file: /Users/bknaepen/Desktop/petsc-2.3.3- >> p6 ((os/kern) >> invalid argument) >> collect2: ld returned 1 exit status >> } >> Pushing language C >> Popping language C >> in mpicc -o conftest -rpath /Users/bknaepen/Desktop/petsc-2.3.3-p6 >> -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress >> -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress >> -fPIC >> conftest.o >> Source: >> #include "confdefs.h" >> #include "conffix.h" >> >> int main() { >> ; >> return 0; >> } >> Rejecting linker flag -rpath /Users/bknaepen/Desktop/ >> petsc-2.3.3-p6 due >> to nonzero status from link >> Rejecting C linker flag -rpath /Users/bknaepen/Desktop/ >> petsc-2.3.3-p6 >> due to >> i686-apple-darwin8-gcc-4.0.1: unrecognized option '-rpath' >> /usr/bin/ld: can't map file: /Users/bknaepen/Desktop/petsc-2.3.3- >> p6 ((os/kern) >> invalid argument) >> collect2: ld returned 1 exit status >> >> Rejected C linker flag -rpath >> Trying C linker flag -Wl,-R, >> sh: mpicc -c -o conftest.o -fPIC conftest.c >> Executing: mpicc -c -o conftest.o -fPIC conftest.c >> sh: >> Pushing language C >> Popping language C >> sh: mpicc -o conftest -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 >> -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress >> -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress >> -fPIC >> conftest.o >> Executing: mpicc -o conftest -Wl,-R,/Users/bknaepen/Desktop/ >> petsc-2.3.3-p6 >> -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress >> -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress >> -fPIC >> conftest.o >> sh: >> Possible ERROR while running linker: /usr/bin/ld: unknown flag: -R >> collect2: ld returned 1 exit status >> output: ret = 256 >> error message = {/usr/bin/ld: unknown flag: -R >> collect2: ld returned 1 exit status >> } >> Pushing language C >> Popping language C >> in mpicc -o conftest -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3-p6 >> -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress >> -Wl,-multiply_defined,suppress -Wl,-multiply_defined -Wl,suppress >> -fPIC >> conftest.o >> Source: >> #include "confdefs.h" >> #include "conffix.h" >> >> int main() { >> ; >> return 0; >> } >> Rejecting linker flag -Wl,-R,/Users/bknaepen/Desktop/ >> petsc-2.3.3-p6 due >> to nonzero status from link >> Rejecting C linker flag -Wl,-R,/Users/bknaepen/Desktop/ >> petsc-2.3.3-p6 >> due to >> /usr/bin/ld: unknown flag: -R >> collect2: ld returned 1 exit status >> >> Rejected C linker flag -Wl,-R, >> Popping language C >> Pushing language FC >> sh: mpicc -flags >> Executing: mpicc -flags >> sh: >> Trying FC linker flag -Wl,-rpath, >> sh: mpif90 -c -o conftest.o -I. -fPIC conftest.F >> Executing: mpif90 -c -o conftest.o -I. -fPIC conftest.F >> sh: >> Pushing language C >> Popping language C >> sh: mpif90 -o conftest -Wl,-rpath,/Users/bknaepen/Desktop/ >> petsc-2.3.3-p6 >> -I. -fPIC conftest.o >> Executing: mpif90 -o conftest >> -Wl,-rpath,/Users/bknaepen/Desktop/petsc-2.3.3-p6 -I. -fPIC >> conftest.o >> sh: >> Possible ERROR while running linker: ipo: error #11019: OBJREAD >> Error: Could >> not create mapping for /Users/bknaepen/Desktop/petsc-2.3.3-p6 >> ifort: error #10014: problem during multi-file optimization >> compilation (code >> 1) >> output: ret = 256 >> error message = {ipo: error #11019: OBJREAD Error: Could not >> create mapping >> for /Users/bknaepen/Desktop/petsc-2.3.3-p6 >> ifort: error #10014: problem during multi-file optimization >> compilation (code >> 1) >> } >> Pushing language C >> Popping language C >> in mpif90 -o conftest -Wl,-rpath,/Users/bknaepen/Desktop/ >> petsc-2.3.3-p6 -I. >> -fPIC conftest.o >> Source: >> program main >> >> end >> Rejecting linker flag -Wl,-rpath,/Users/bknaepen/Desktop/ >> petsc-2.3.3-p6 >> due to nonzero status from link >> Valid FC linker flag -Wl,-rpath,/Users/bknaepen/Desktop/ >> petsc-2.3.3-p6 >> Rejected FC linker flag -Wl,-rpath, >> Trying FC linker flag -R >> sh: mpif90 -c -o conftest.o -I. -fPIC conftest.F >> Executing: mpif90 -c -o conftest.o -I. -fPIC conftest.F >> sh: >> Pushing language C >> Popping language C >> sh: mpif90 -o conftest -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 >> -I. -fPIC >> conftest.o >> Executing: mpif90 -o conftest -R/Users/bknaepen/Desktop/ >> petsc-2.3.3-p6 -I. >> -fPIC conftest.o >> sh: >> Possible ERROR while running linker: ifort: command line warning >> #10006: >> ignoring unknown option '-R/Users/bknaepen/Desktop/petsc-2.3.3-p6' >> output: error message = {ifort: command line warning #10006: >> ignoring unknown >> option '-R/Users/bknaepen/Desktop/petsc-2.3.3-p6' >> } >> Pushing language C >> Popping language C >> in mpif90 -o conftest -R/Users/bknaepen/Desktop/petsc-2.3.3-p6 - >> I. -fPIC >> conftest.o >> Source: >> program main >> >> end >> Rejecting FC linker flag -R/Users/bknaepen/Desktop/ >> petsc-2.3.3-p6 due >> to >> ifort: command line warning #10006: ignoring unknown option >> '-R/Users/bknaepen/Desktop/petsc-2.3.3-p6' >> >> Rejected FC linker flag -R >> Trying FC linker flag -rpath >> sh: mpif90 -c -o conftest.o -I. -fPIC conftest.F >> Executing: mpif90 -c -o conftest.o -I. -fPIC conftest.F >> sh: >> Pushing language C >> Popping language C >> sh: mpif90 -o conftest -rpath /Users/bknaepen/Desktop/ >> petsc-2.3.3-p6 -I. >> -fPIC conftest.o >> Executing: mpif90 -o conftest -rpath /Users/bknaepen/Desktop/ >> petsc-2.3.3-p6 >> -I. -fPIC conftest.o >> Runaway process exceeded time limit of 600.0s >> Possible ERROR while running linker: Runaway process exceeded time >> limit of >> 600.0s >> output: Runaway processret = -1 >> error message = {Runaway process exceeded time limit of 600.0s >> } >> Pushing language C >> Popping language C >> in mpif90 -o conftest -rpath /Users/bknaepen/Desktop/petsc-2.3.3- >> p6 -I. >> -fPIC conftest.o >> Source: >> program main >> >> end >> Rejecting linker flag -rpath /Users/bknaepen/Desktop/ >> petsc-2.3.3-p6 due >> to nonzero status from link >> Valid FC linker flag -rpath /Users/bknaepen/Desktop/ >> petsc-2.3.3-p6 >> Rejected FC linker flag -rpath >> Trying FC linker flag -Wl,-R, >> sh: mpif90 -c -o conftest.o -I. -fPIC conftest.F >> Executing: mpif90 -c -o conftest.o -I. -fPIC conftest.F >> sh: >> Pushing language C >> Popping language C >> sh: mpif90 -o conftest -Wl,-R,/Users/bknaepen/Desktop/ >> petsc-2.3.3-p6 -I. >> -fPIC conftest.o >> Executing: mpif90 -o conftest -Wl,-R,/Users/bknaepen/Desktop/ >> petsc-2.3.3-p6 >> -I. -fPIC conftest.o >> sh: >> Possible ERROR while running linker: IPO link: Warning unknown >> option '-R'. >> ipo: error #11019: OBJREAD Error: Could not create mapping for >> /Users/bknaepen/Desktop/petsc-2.3.3-p6 >> ifort: error #10014: problem during multi-file optimization >> compilation (code >> 1) >> output: ret = 256 >> error message = {IPO link: Warning unknown option '-R'. >> ipo: error #11019: OBJREAD Error: Could not create mapping for >> /Users/bknaepen/Desktop/petsc-2.3.3-p6 >> ifort: error #10014: problem during multi-file optimization >> compilation (code >> 1) >> } >> Pushing language C >> Popping language C >> in mpif90 -o conftest -Wl,-R,/Users/bknaepen/Desktop/petsc-2.3.3- >> p6 -I. >> -fPIC conftest.o >> Source: >> program main >> >> end >> Rejecting linker flag -Wl,-R,/Users/bknaepen/Desktop/ >> petsc-2.3.3-p6 due >> to nonzero status from link >> Rejecting FC linker flag -Wl,-R,/Users/bknaepen/Desktop/ >> petsc-2.3.3-p6 >> due to >> IPO link: Warning unknown option '-R'. >> ipo: error #11019: OBJREAD Error: Could not create mapping for >> /Users/bknaepen/Desktop/petsc-2.3.3-p6 >> ifort: error #10014: problem during multi-file optimization >> compilation (code >> 1) >> >> Rejected FC linker flag -Wl,-R, >> Popping language FC >> ===================================================================== >> =========== >> On 12 Oct 2007, at 23:26, Barry Smith wrote: >> >>> >>> I have run with 10.0.0.20 on the Mac with no difficulties using >>> the options >>> below with petsc-2.3.3-6 and have no hanging. config/configure.py >>> just takes >>> a >>> couple of minutes. >>> >>> [bsmith:~/Src/petsc-release-2.3.3] bsmith% uname -a >>> Darwin bsmith.mcs.anl.gov 8.10.1 Darwin Kernel Version 8.10.1: >>> Wed May 23 >>> 16:33:00 PDT 2007; root:xnu-792.22.5~1/RELEASE_I386 i386 i386 >>> >>> >>> Barry >>> >>> >>> On Fri, 12 Oct 2007, Knaepen Bernard wrote: >>> >>>> Hello, >>>> >>>> is it normal that the ./config/configure.py stalls for a long >>>> time when >>>> building petsc-2.3.3-p6 on MacOSX? The step at which it stays >>>> stuck is: >>>> >>>> TESTING: checkSharedLinkerPaths from >>>> config.setCompilers(python/BuildSystem/config/setCompilers.py:1107) >>>> >>>> It remains there for around five minutes then continues the >>>> configure >>>> process without problem. Except that it leaves a fortran process >>>> alive >>>> that I have to manually kill later on. This process looks like: >>>> >>>> 17623 fortcom 60.8% 5:06.44 1 13 60 364K 12.1M >>>> 2.23M >>>> 57.8M >>>> 17614 ifortia32 0.0% 0:00.02 1 14 19 392K 1.96M >>>> 1.86M >>>> 29.0M >>>> 17613 ifortbin 0.0% 0:00.01 1 13 16 148K 992K >>>> 876K >>>> 28.6M >>>> >>>> when I use the top command. >>>> >>>> I am configuring with the following options: >>>> >>>> /config/configure.py --with-cc=mpicc --with-fc=mpif90 --with- >>>> hypre=1 >>>> --download-hypre=1 --with-debugging=0 COPTFLAGS=-O3 FOPTFLAGS=-O3 >>>> --with-cxx=mpicxx --with-x=0 --with-sudo=sudo >>>> >>>> The system is MacOSX 10.4.9 (latest) running on a MacPro (or >>>> Macbook Pro) >>>> and the fortran compiler is ifort 10.0.17. >>>> >>>> Regards, >>>> Bernard. >>>> >>>> >>>> >>>> >> From jwicks at cs.brown.edu Mon Oct 15 12:16:01 2007 From: jwicks at cs.brown.edu (John R. Wicks) Date: Mon, 15 Oct 2007 13:16:01 -0400 Subject: Permissions problem Message-ID: <000801c80f4f$113fb540$6b01a8c0@jwickslptp> I am running petsc-2.3.3-p3 on two platforms (Debian GNU/Linux 4.0 and Mac OS Server 10.4.10) installed in a shared NFS directory. While I can get mpdboot to run on both platforms, and I can run multi-machine programs using mpiexec under Linux, when I try to do it on the Mac, I get a permissions error: newton ~/pweb/C % mpiexec -n 5 t problem with execution of t on poinsot.ilab.cs.brown.edu: [Errno 13] Permission denied problem with execution of t on pell.ilab.cs.brown.edu: [Errno 13] Permission denied problem with execution of t on newton.ilab.cs.brown.edu: [Errno 13] Permission denied problem with execution of t on pell.ilab.cs.brown.edu: [Errno 13] Permission denied problem with execution of t on newton.ilab.cs.brown.edu: [Errno 13] Permission denied What permissions does this need? From balay at mcs.anl.gov Mon Oct 15 13:16:25 2007 From: balay at mcs.anl.gov (Satish Balay) Date: Mon, 15 Oct 2007 13:16:25 -0500 (CDT) Subject: Permissions problem In-Reply-To: <000801c80f4f$113fb540$6b01a8c0@jwickslptp> References: <000801c80f4f$113fb540$6b01a8c0@jwickslptp> Message-ID: Things to check. run a basic MPI example mpiexec -n 5 ./t ssh poinsot.ilab.cs.brown.edu `pwd`/t Satish On Mon, 15 Oct 2007, John R. Wicks wrote: > I am running petsc-2.3.3-p3 on two platforms (Debian GNU/Linux 4.0 and Mac > OS Server 10.4.10) installed in a shared NFS directory. While I can get > mpdboot to run on both platforms, and I can run multi-machine programs using > mpiexec under Linux, when I try to do it on the Mac, I get a permissions > error: > > newton ~/pweb/C % mpiexec -n 5 t > problem with execution of t on poinsot.ilab.cs.brown.edu: [Errno 13] > Permission denied > problem with execution of t on pell.ilab.cs.brown.edu: [Errno 13] > Permission denied > problem with execution of t on newton.ilab.cs.brown.edu: [Errno 13] > Permission denied > problem with execution of t on pell.ilab.cs.brown.edu: [Errno 13] > Permission denied > problem with execution of t on newton.ilab.cs.brown.edu: [Errno 13] > Permission denied > > What permissions does this need? > > From tyoung at ippt.gov.pl Wed Oct 17 02:49:21 2007 From: tyoung at ippt.gov.pl (Toby D. Young) Date: Wed, 17 Oct 2007 09:49:21 +0200 Subject: More wierd behaviour from MatIsSymmetric? Message-ID: <20071017094921.3dbbf6f1@rav.ippt.gov.pl> I am having a little trouble, or confusion, with the MatIsSymmetric function. I am getting the reverse behaviour from this routine that I expect. I am convinced I have a symmetric matrix, but the answer I get from this function is always zero - ie, PetscTruth = PetscFalse. Why, if PETSC_YES is an alias for PETSC_TRUE, do they have opposite behaviour according to the documentation? That doen't seem right. What is the idea here? I there an error in the documentation, or am I misunderstanding something? I appreceate your usually helpful remarks!! For reference, please see: http://www-unix.mcs.anl.gov/petsc/petsc-as/snapshots/petsc-current/include/petsc.h.html#PetscTruth Or the notes written below. Using petsc 2.3.0; Best, Toby 339: /*MC 340: PETSC_TRUE - True value of PetscTruth 342: Level: beginner 344: Note: Nonzero integer 346: .seealso: PetscTruth, PETSC_FALSE 347: M*/ 349: /*MC 350: PETSC_YES - Alias for PETSC_TRUE 352: Level: beginner 354: Note: Zero integer 356: .seealso: PetscTruth, PETSC_TRUE, PETSC_FALSE, PETSC_NO 357: M*/ -- Toby D. Young - Adiunkt (Assistant Professor) Department of Computational Science Institute of Fundamental Technological Research Polish Academy of Sciences Room 206, Swietokrzyska 21 00-049 Warsaw, POLAND From ba001ing at unibas.it Wed Oct 17 04:40:43 2007 From: ba001ing at unibas.it (Aldo Bonfiglioli) Date: Wed, 17 Oct 2007 11:40:43 +0200 Subject: Non repeatability issue Message-ID: <4715D89B.8070005@unibas.it> Dear users, in the notes attached here we address an issue concerning nonrepeatability. We know this is a known issue in parallel floating point programs, but we would like to be sure we are interpreting things correctly a not mis-using PETSc. Comments are welcome. Regards, Aldo -- Dr. Aldo Bonfiglioli Dip.to di Ingegneria e Fisica dell'Ambiente (DIFA) Universita' della Basilicata V.le dell'Ateneo lucano, 10 85100 Potenza ITALY tel:+39.0971.205203 fax:+39.0971.205160 -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: README.nonrepeatibility URL: From knepley at gmail.com Wed Oct 17 18:20:18 2007 From: knepley at gmail.com (Matthew Knepley) Date: Wed, 17 Oct 2007 18:20:18 -0500 Subject: More wierd behaviour from MatIsSymmetric? In-Reply-To: <20071017094921.3dbbf6f1@rav.ippt.gov.pl> References: <20071017094921.3dbbf6f1@rav.ippt.gov.pl> Message-ID: Is there a reason for not using 2.3.3? I suspect this to be a problem with finite precision, which is corrected in the newest release.' Thanks, Matt On 10/17/07, Toby D. Young wrote: > > > I am having a little trouble, or confusion, with the MatIsSymmetric > function. I am getting the reverse behaviour from this routine that I > expect. I am convinced I have a symmetric matrix, but the answer I get > from this function is always zero - ie, PetscTruth = PetscFalse. > > Why, if PETSC_YES is an alias for PETSC_TRUE, do they have opposite > behaviour according to the documentation? That doen't seem right. What > is the idea here? > > I there an error in the documentation, or am I misunderstanding > something? I appreceate your usually helpful remarks!! > > For reference, please see: > http://www-unix.mcs.anl.gov/petsc/petsc-as/snapshots/petsc-current/include/petsc.h.html#PetscTruth > Or the notes written below. > > Using petsc 2.3.0; > Best, > Toby > > > 339: /*MC > 340: PETSC_TRUE - True value of PetscTruth > > 342: Level: beginner > > 344: Note: Nonzero integer > > 346: .seealso: PetscTruth, PETSC_FALSE > 347: M*/ > > 349: /*MC > 350: PETSC_YES - Alias for PETSC_TRUE > > 352: Level: beginner > > 354: Note: Zero integer > > 356: .seealso: PetscTruth, PETSC_TRUE, PETSC_FALSE, PETSC_NO > 357: M*/ > > > -- > > Toby D. Young - Adiunkt (Assistant Professor) > Department of Computational Science > Institute of Fundamental Technological Research > Polish Academy of Sciences > Room 206, Swietokrzyska 21 > 00-049 Warsaw, POLAND > > -- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. -- Norbert Wiener From bsmith at mcs.anl.gov Wed Oct 17 22:06:04 2007 From: bsmith at mcs.anl.gov (Barry Smith) Date: Wed, 17 Oct 2007 22:06:04 -0500 (CDT) Subject: Non repeatability issue In-Reply-To: <4715D89B.8070005@unibas.it> References: <4715D89B.8070005@unibas.it> Message-ID: The 2d results, as noted, are reasonable and expected behavior. The 3d results are not acceptable, when sometimes it converges and sometimes it does not. I would use a 1e-12 KSP rtol and a similar 1e-12 SNES rtol; does it now always converge? If not then I suspect there may be some error in the residual (function) computation. Barry On Wed, 17 Oct 2007, Aldo Bonfiglioli wrote: > Dear users, > in the notes attached here we address an issue concerning > nonrepeatability. > We know this is a known issue in parallel floating point programs, > but we would like to be sure we are interpreting > things correctly a not mis-using PETSc. > > Comments are welcome. > > Regards, > Aldo > > From timothy.stitt at ichec.ie Mon Oct 22 15:11:31 2007 From: timothy.stitt at ichec.ie (Tim Stitt) Date: Mon, 22 Oct 2007 21:11:31 +0100 Subject: MatLUFactorSymbolic Message-ID: <471D03F3.7070700@ichec.ie> Hi all, I am trying to call the routine MatLUFactorSymbolic in my Fortran PETSc code and I am having some difficulties in figuring out what argument should be passed to the MatFactorInfo *info parameter? Can someone describe the possible values that this parameter can take within a Fortran calling interface? Thanks in advance, Tim. From balay at mcs.anl.gov Mon Oct 22 15:18:53 2007 From: balay at mcs.anl.gov (Satish Balay) Date: Mon, 22 Oct 2007 15:18:53 -0500 (CDT) Subject: MatLUFactorSymbolic In-Reply-To: <471D03F3.7070700@ichec.ie> References: <471D03F3.7070700@ichec.ie> Message-ID: The usage of MatFactorInfo is similar to MatInfo. An example usage of MatInfo is at src/ksp/ksp/examples/tests/ex12f.F And you would need to look at include/finclude/petscmat.h for the different flags that can be set in MatFactorInfo Satish On Mon, 22 Oct 2007, Tim Stitt wrote: > Hi all, > > I am trying to call the routine MatLUFactorSymbolic in my Fortran PETSc code > and I am having some difficulties in figuring out what argument should be > passed to the > > MatFactorInfo *info > > parameter? > > Can someone describe the possible values that this parameter can take within a > Fortran calling interface? > > Thanks in advance, > > Tim. > > From bsmith at mcs.anl.gov Mon Oct 22 15:25:15 2007 From: bsmith at mcs.anl.gov (Barry Smith) Date: Mon, 22 Oct 2007 15:25:15 -0500 (CDT) Subject: MatLUFactorSymbolic In-Reply-To: <471D03F3.7070700@ichec.ie> References: <471D03F3.7070700@ichec.ie> Message-ID: It should be MatFactorInfo info(MAT_FACTORINFO_SIZE) The elements in the info array can found at http://www-unix.mcs.anl.gov/petsc/petsc-as/snapshots/petsc-current/docs/manualpages/Mat/MatFactorInfo.html You can call MatFactorInfoInitialize() to set default values into the array and then change those you wish to (or change none). Barry I have added this info to the MatFactorInfo manual page. On Mon, 22 Oct 2007, Tim Stitt wrote: > Hi all, > > I am trying to call the routine MatLUFactorSymbolic in my Fortran PETSc code > and I am having some difficulties in figuring out what argument should be > passed to the > > MatFactorInfo *info > > parameter? > > Can someone describe the possible values that this parameter can take within a > Fortran calling interface? > > Thanks in advance, > > Tim. > > From timothy.stitt at ichec.ie Mon Oct 22 15:41:49 2007 From: timothy.stitt at ichec.ie (Tim Stitt) Date: Mon, 22 Oct 2007 21:41:49 +0100 Subject: MatLUFactorSymbolic In-Reply-To: References: <471D03F3.7070700@ichec.ie> Message-ID: <471D0B0D.5000902@ichec.ie> Many thanks...that helps immensely. Barry Smith wrote: > It should be > > MatFactorInfo info(MAT_FACTORINFO_SIZE) > > The elements in the info array can found at > http://www-unix.mcs.anl.gov/petsc/petsc-as/snapshots/petsc-current/docs/manualpages/Mat/MatFactorInfo.html > You can call MatFactorInfoInitialize() to set default values into the array and > then change those you wish to (or change none). > > Barry > > I have added this info to the MatFactorInfo manual page. > > > On Mon, 22 Oct 2007, Tim Stitt wrote: > > >> Hi all, >> >> I am trying to call the routine MatLUFactorSymbolic in my Fortran PETSc code >> and I am having some difficulties in figuring out what argument should be >> passed to the >> >> MatFactorInfo *info >> >> parameter? >> >> Can someone describe the possible values that this parameter can take within a >> Fortran calling interface? >> >> Thanks in advance, >> >> Tim. >> >> >> > > From bsmith at mcs.anl.gov Tue Oct 23 16:22:39 2007 From: bsmith at mcs.anl.gov (Barry Smith) Date: Tue, 23 Oct 2007 16:22:39 -0500 (CDT) Subject: ISLocalToGlobalMappingApply(mapping,m,input,output); and alike "functions" In-Reply-To: References: Message-ID: This is a known problem. I think the correct solution is to replace all the "macro functions" with PETSC_STATIC_INLINE functions. I have done this in petsc-dev http://www-unix.mcs.anl.gov/petsc/petsc-as/developers/index.html Barry On Thu, 11 Oct 2007, Mads Hoel wrote: > Hi all, > > I been having trouble with this function today. I called it from a function > returning a void. Later found out it was actually > a macro with a "return > PetscError(__LINE__,__FUNCT__,__FILE__,__SDIR__,n,1,s,a1,a2,a3);" statement in > it. Which makes it > not possible to call ISLocalToGlobalMappingApply from a function returning a > void, when compiling with g++ compiler. > Or perhaps it may lead to difficult to find error in functions returning > another type than an int, for example an object with > a constructor accepting an int. > > Thanks > > From z.sheng at ewi.tudelft.nl Wed Oct 24 08:50:27 2007 From: z.sheng at ewi.tudelft.nl (Zhifeng Sheng) Date: Wed, 24 Oct 2007 15:50:27 +0200 Subject: Validate a Object Message-ID: <471F4DA3.3070204@ewi.tudelft.nl> Dear all I am programming a simulator with PetSc, and I need to create matrix and vector which shall be destoried in the end. However, the problem is : sometimes, a matrix is declared but never created, and if I try to destory this object, Petsc would dump error. Is there anyway to check whether a Mat or Vec has been created or not? e.g Mat A; MatDestory A; // this will dump error, which function should I use to check whether A is created? Thank you. Best regards Zhifeng Sheng From balay at mcs.anl.gov Wed Oct 24 09:10:40 2007 From: balay at mcs.anl.gov (Satish Balay) Date: Wed, 24 Oct 2007 09:10:40 -0500 (CDT) Subject: Validate a Object In-Reply-To: <471F4DA3.3070204@ewi.tudelft.nl> References: <471F4DA3.3070204@ewi.tudelft.nl> Message-ID: Mat A=0; if (A) MatDestroy(A); Satish On Wed, 24 Oct 2007, Zhifeng Sheng wrote: > Dear all > > I am programming a simulator with PetSc, and I need to create matrix and > vector which shall be destoried in the end. > > However, the problem is : sometimes, a matrix is declared but never created, > and if I try to destory this object, Petsc would dump error. > > Is there anyway to check whether a Mat or Vec has been created or not? > > > e.g > > > Mat A; > MatDestory A; // this will dump error, which function should I use to check > whether A is created? > > Thank you. > Best regards > Zhifeng Sheng > > From timothy.stitt at ichec.ie Thu Oct 25 13:54:20 2007 From: timothy.stitt at ichec.ie (Tim Stitt) Date: Thu, 25 Oct 2007 19:54:20 +0100 Subject: Invalid argument! - not a permutation! Message-ID: <4720E65C.9070504@ichec.ie> Hi all, Following on from some help I got a few days ago with respect to MatLUFactorSymbolic() usage in Fortran...I think I now have all the parts in place but keep getting the following runtime error upon execution: [0]PETSC ERROR: --------------------- Error Message ------------------------------------ [0]PETSC ERROR: Invalid argument! [0]PETSC ERROR: not a permutation! ... [0]PETSC ERROR: ------------------------------------------------------------------------ [0]PETSC ERROR: ISInvertPermutation() line 183 in src/vec/is/interface/index.c [0]PETSC ERROR: MatLUFactorSymbolic_SeqAIJ() line 276 in src/mat/impls/aij/seq/aijfact.c [0]PETSC ERROR: MatLUFactorSymbolic() line 2178 in src/mat/interface/matrix.c [localhost.localdomain:24730] MPI_ABORT invoked on rank 0 in communicator MPI_COMM_WORLD with errorcode 62 Here is the portion of the offending code (I think): call MatAssemblyBegin(A,MAT_FINAL_ASSEMBLY,error);CHKERRQ(error) call MatAssemblyEnd(A,MAT_FINAL_ASSEMBLY,error);CHKERRQ(error) call ISCreateGeneral(PETSC_COMM_WORLD,order,columnIndices, indexSet,error);CHKERRQ(error) call MatFactorInfoInitialize(info,error);CHKERRQ(error) call MatLUFactorSymbolic(A,indexSet,indexSet,info,factorMat,error);CHKERRQ(error) order - number of rows/columns in square matrix A columnIndices - 1D array containing values 0-->order-1 All types are set correctly as far as I can tell. Can anyone suggest why the execution falls over during the MatLUFactorSymbolic() call? Thanks again, Tim. -- Dr. Timothy Stitt HPC Application Consultant - ICHEC (www.ichec.ie) Dublin Institute for Advanced Studies 5 Merrion Square - Dublin 2 - Ireland +353-1-6621333 (tel) / +353-1-6621477 (fax) From knepley at gmail.com Thu Oct 25 14:00:09 2007 From: knepley at gmail.com (Matthew Knepley) Date: Thu, 25 Oct 2007 14:00:09 -0500 Subject: Invalid argument! - not a permutation! In-Reply-To: <4720E65C.9070504@ichec.ie> References: <4720E65C.9070504@ichec.ie> Message-ID: I think you need to tell use that the IS is a permutation: http://www-unix.mcs.anl.gov/petsc/petsc-as/snapshots/petsc-current/docs/manualpages/IS/ISSetPermutation.html Matt On 10/25/07, Tim Stitt wrote: > Hi all, > > Following on from some help I got a few days ago with respect to > MatLUFactorSymbolic() usage in Fortran...I think I now have all the > parts in place but keep getting the following runtime error upon execution: > > [0]PETSC ERROR: --------------------- Error Message > ------------------------------------ > [0]PETSC ERROR: Invalid argument! > [0]PETSC ERROR: not a permutation! > ... > [0]PETSC ERROR: > ------------------------------------------------------------------------ > [0]PETSC ERROR: ISInvertPermutation() line 183 in > src/vec/is/interface/index.c > [0]PETSC ERROR: MatLUFactorSymbolic_SeqAIJ() line 276 in > src/mat/impls/aij/seq/aijfact.c > [0]PETSC ERROR: MatLUFactorSymbolic() line 2178 in > src/mat/interface/matrix.c > [localhost.localdomain:24730] MPI_ABORT invoked on rank 0 in > communicator MPI_COMM_WORLD with errorcode 62 > > Here is the portion of the offending code (I think): > > call MatAssemblyBegin(A,MAT_FINAL_ASSEMBLY,error);CHKERRQ(error) > call MatAssemblyEnd(A,MAT_FINAL_ASSEMBLY,error);CHKERRQ(error) > call ISCreateGeneral(PETSC_COMM_WORLD,order,columnIndices, > indexSet,error);CHKERRQ(error) > call MatFactorInfoInitialize(info,error);CHKERRQ(error) > call > MatLUFactorSymbolic(A,indexSet,indexSet,info,factorMat,error);CHKERRQ(error) > > order - number of rows/columns in square matrix A > columnIndices - 1D array containing values 0-->order-1 > > All types are set correctly as far as I can tell. > > Can anyone suggest why the execution falls over during the > MatLUFactorSymbolic() call? > > Thanks again, > > Tim. > > -- > Dr. Timothy Stitt > HPC Application Consultant - ICHEC (www.ichec.ie) > > Dublin Institute for Advanced Studies > 5 Merrion Square - Dublin 2 - Ireland > > +353-1-6621333 (tel) / +353-1-6621477 (fax) > > -- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. -- Norbert Wiener From bsmith at mcs.anl.gov Thu Oct 25 14:04:09 2007 From: bsmith at mcs.anl.gov (Barry Smith) Date: Thu, 25 Oct 2007 14:04:09 -0500 (CDT) Subject: Invalid argument! - not a permutation! In-Reply-To: References: <4720E65C.9070504@ichec.ie> Message-ID: I have added this information to the error message that gets printed. Barry On Thu, 25 Oct 2007, Matthew Knepley wrote: > I think you need to tell use that the IS is a permutation: > > http://www-unix.mcs.anl.gov/petsc/petsc-as/snapshots/petsc-current/docs/manualpages/IS/ISSetPermutation.html > > Matt > > On 10/25/07, Tim Stitt wrote: > > Hi all, > > > > Following on from some help I got a few days ago with respect to > > MatLUFactorSymbolic() usage in Fortran...I think I now have all the > > parts in place but keep getting the following runtime error upon execution: > > > > [0]PETSC ERROR: --------------------- Error Message > > ------------------------------------ > > [0]PETSC ERROR: Invalid argument! > > [0]PETSC ERROR: not a permutation! > > ... > > [0]PETSC ERROR: > > ------------------------------------------------------------------------ > > [0]PETSC ERROR: ISInvertPermutation() line 183 in > > src/vec/is/interface/index.c > > [0]PETSC ERROR: MatLUFactorSymbolic_SeqAIJ() line 276 in > > src/mat/impls/aij/seq/aijfact.c > > [0]PETSC ERROR: MatLUFactorSymbolic() line 2178 in > > src/mat/interface/matrix.c > > [localhost.localdomain:24730] MPI_ABORT invoked on rank 0 in > > communicator MPI_COMM_WORLD with errorcode 62 > > > > Here is the portion of the offending code (I think): > > > > call MatAssemblyBegin(A,MAT_FINAL_ASSEMBLY,error);CHKERRQ(error) > > call MatAssemblyEnd(A,MAT_FINAL_ASSEMBLY,error);CHKERRQ(error) > > call ISCreateGeneral(PETSC_COMM_WORLD,order,columnIndices, > > indexSet,error);CHKERRQ(error) > > call MatFactorInfoInitialize(info,error);CHKERRQ(error) > > call > > MatLUFactorSymbolic(A,indexSet,indexSet,info,factorMat,error);CHKERRQ(error) > > > > order - number of rows/columns in square matrix A > > columnIndices - 1D array containing values 0-->order-1 > > > > All types are set correctly as far as I can tell. > > > > Can anyone suggest why the execution falls over during the > > MatLUFactorSymbolic() call? > > > > Thanks again, > > > > Tim. > > > > -- > > Dr. Timothy Stitt > > HPC Application Consultant - ICHEC (www.ichec.ie) > > > > Dublin Institute for Advanced Studies > > 5 Merrion Square - Dublin 2 - Ireland > > > > +353-1-6621333 (tel) / +353-1-6621477 (fax) > > > > > > > From timothy.stitt at ichec.ie Thu Oct 25 14:11:13 2007 From: timothy.stitt at ichec.ie (Tim Stitt) Date: Thu, 25 Oct 2007 20:11:13 +0100 Subject: Invalid argument! - not a permutation! In-Reply-To: References: <4720E65C.9070504@ichec.ie> Message-ID: <4720EA51.7080907@ichec.ie> Cheers Matthew...that worked now. Matthew Knepley wrote: > I think you need to tell use that the IS is a permutation: > > http://www-unix.mcs.anl.gov/petsc/petsc-as/snapshots/petsc-current/docs/manualpages/IS/ISSetPermutation.html > > Matt > > On 10/25/07, Tim Stitt wrote: > >> Hi all, >> >> Following on from some help I got a few days ago with respect to >> MatLUFactorSymbolic() usage in Fortran...I think I now have all the >> parts in place but keep getting the following runtime error upon execution: >> >> [0]PETSC ERROR: --------------------- Error Message >> ------------------------------------ >> [0]PETSC ERROR: Invalid argument! >> [0]PETSC ERROR: not a permutation! >> ... >> [0]PETSC ERROR: >> ------------------------------------------------------------------------ >> [0]PETSC ERROR: ISInvertPermutation() line 183 in >> src/vec/is/interface/index.c >> [0]PETSC ERROR: MatLUFactorSymbolic_SeqAIJ() line 276 in >> src/mat/impls/aij/seq/aijfact.c >> [0]PETSC ERROR: MatLUFactorSymbolic() line 2178 in >> src/mat/interface/matrix.c >> [localhost.localdomain:24730] MPI_ABORT invoked on rank 0 in >> communicator MPI_COMM_WORLD with errorcode 62 >> >> Here is the portion of the offending code (I think): >> >> call MatAssemblyBegin(A,MAT_FINAL_ASSEMBLY,error);CHKERRQ(error) >> call MatAssemblyEnd(A,MAT_FINAL_ASSEMBLY,error);CHKERRQ(error) >> call ISCreateGeneral(PETSC_COMM_WORLD,order,columnIndices, >> indexSet,error);CHKERRQ(error) >> call MatFactorInfoInitialize(info,error);CHKERRQ(error) >> call >> MatLUFactorSymbolic(A,indexSet,indexSet,info,factorMat,error);CHKERRQ(error) >> >> order - number of rows/columns in square matrix A >> columnIndices - 1D array containing values 0-->order-1 >> >> All types are set correctly as far as I can tell. >> >> Can anyone suggest why the execution falls over during the >> MatLUFactorSymbolic() call? >> >> Thanks again, >> >> Tim. >> >> -- >> Dr. Timothy Stitt >> HPC Application Consultant - ICHEC (www.ichec.ie) >> >> Dublin Institute for Advanced Studies >> 5 Merrion Square - Dublin 2 - Ireland >> >> +353-1-6621333 (tel) / +353-1-6621477 (fax) >> >> >> > > > -- Dr. Timothy Stitt HPC Application Consultant - ICHEC (www.ichec.ie) Dublin Institute for Advanced Studies 5 Merrion Square - Dublin 2 - Ireland +353-1-6621333 (tel) / +353-1-6621477 (fax) From timothy.stitt at ichec.ie Thu Oct 25 14:22:58 2007 From: timothy.stitt at ichec.ie (Tim Stitt) Date: Thu, 25 Oct 2007 20:22:58 +0100 Subject: Invalid argument! - not a permutation! In-Reply-To: References: <4720E65C.9070504@ichec.ie> Message-ID: <4720ED12.4040708@ichec.ie> OK...that is useful thanks. If you don't mind just one more question...I notice that the MatLUFactorSymbolic() routine now executes with one processor but complains when I run > 2 processors. The actual error message is: [1]PETSC ERROR: No support for this operation for this object type! [1]PETSC ERROR: Matrix type mpiaij symbolic LU! Is this because the default LU factorisation routine isn't parallel...I just assumed parallel superlu would be picked up automatically (it is compiled into my PETSc build) ? Tim. Barry Smith wrote: > I have added this information to the error message that gets printed. > > Barry > > > On Thu, 25 Oct 2007, Matthew Knepley wrote: > > >> I think you need to tell use that the IS is a permutation: >> >> http://www-unix.mcs.anl.gov/petsc/petsc-as/snapshots/petsc-current/docs/manualpages/IS/ISSetPermutation.html >> >> Matt >> >> On 10/25/07, Tim Stitt wrote: >> >>> Hi all, >>> >>> Following on from some help I got a few days ago with respect to >>> MatLUFactorSymbolic() usage in Fortran...I think I now have all the >>> parts in place but keep getting the following runtime error upon execution: >>> >>> [0]PETSC ERROR: --------------------- Error Message >>> ------------------------------------ >>> [0]PETSC ERROR: Invalid argument! >>> [0]PETSC ERROR: not a permutation! >>> ... >>> [0]PETSC ERROR: >>> ------------------------------------------------------------------------ >>> [0]PETSC ERROR: ISInvertPermutation() line 183 in >>> src/vec/is/interface/index.c >>> [0]PETSC ERROR: MatLUFactorSymbolic_SeqAIJ() line 276 in >>> src/mat/impls/aij/seq/aijfact.c >>> [0]PETSC ERROR: MatLUFactorSymbolic() line 2178 in >>> src/mat/interface/matrix.c >>> [localhost.localdomain:24730] MPI_ABORT invoked on rank 0 in >>> communicator MPI_COMM_WORLD with errorcode 62 >>> >>> Here is the portion of the offending code (I think): >>> >>> call MatAssemblyBegin(A,MAT_FINAL_ASSEMBLY,error);CHKERRQ(error) >>> call MatAssemblyEnd(A,MAT_FINAL_ASSEMBLY,error);CHKERRQ(error) >>> call ISCreateGeneral(PETSC_COMM_WORLD,order,columnIndices, >>> indexSet,error);CHKERRQ(error) >>> call MatFactorInfoInitialize(info,error);CHKERRQ(error) >>> call >>> MatLUFactorSymbolic(A,indexSet,indexSet,info,factorMat,error);CHKERRQ(error) >>> >>> order - number of rows/columns in square matrix A >>> columnIndices - 1D array containing values 0-->order-1 >>> >>> All types are set correctly as far as I can tell. >>> >>> Can anyone suggest why the execution falls over during the >>> MatLUFactorSymbolic() call? >>> >>> Thanks again, >>> >>> Tim. >>> >>> -- >>> Dr. Timothy Stitt >>> HPC Application Consultant - ICHEC (www.ichec.ie) >>> >>> Dublin Institute for Advanced Studies >>> 5 Merrion Square - Dublin 2 - Ireland >>> >>> +353-1-6621333 (tel) / +353-1-6621477 (fax) >>> >>> >>> >> >> > > -- Dr. Timothy Stitt HPC Application Consultant - ICHEC (www.ichec.ie) Dublin Institute for Advanced Studies 5 Merrion Square - Dublin 2 - Ireland +353-1-6621333 (tel) / +353-1-6621477 (fax) From knepley at gmail.com Thu Oct 25 14:26:12 2007 From: knepley at gmail.com (Matthew Knepley) Date: Thu, 25 Oct 2007 14:26:12 -0500 Subject: Invalid argument! - not a permutation! In-Reply-To: <4720ED12.4040708@ichec.ie> References: <4720E65C.9070504@ichec.ie> <4720ED12.4040708@ichec.ie> Message-ID: On 10/25/07, Tim Stitt wrote: > OK...that is useful thanks. > > If you don't mind just one more question...I notice that the > MatLUFactorSymbolic() routine now executes with one processor but > complains when I run > 2 processors. The actual error message is: > > [1]PETSC ERROR: No support for this operation for this object type! > [1]PETSC ERROR: Matrix type mpiaij symbolic LU! > > Is this because the default LU factorisation routine isn't parallel...I > just assumed parallel superlu would be picked up automatically (it is > compiled into my PETSc build) ? It is only valid for the superlu matrix type, -mat_type superlu_dist. Matt > Tim. > > Barry Smith wrote: > > I have added this information to the error message that gets printed. > > > > Barry > > > > > > On Thu, 25 Oct 2007, Matthew Knepley wrote: > > > > > >> I think you need to tell use that the IS is a permutation: > >> > >> http://www-unix.mcs.anl.gov/petsc/petsc-as/snapshots/petsc-current/docs/manualpages/IS/ISSetPermutation.html > >> > >> Matt > >> > >> On 10/25/07, Tim Stitt wrote: > >> > >>> Hi all, > >>> > >>> Following on from some help I got a few days ago with respect to > >>> MatLUFactorSymbolic() usage in Fortran...I think I now have all the > >>> parts in place but keep getting the following runtime error upon execution: > >>> > >>> [0]PETSC ERROR: --------------------- Error Message > >>> ------------------------------------ > >>> [0]PETSC ERROR: Invalid argument! > >>> [0]PETSC ERROR: not a permutation! > >>> ... > >>> [0]PETSC ERROR: > >>> ------------------------------------------------------------------------ > >>> [0]PETSC ERROR: ISInvertPermutation() line 183 in > >>> src/vec/is/interface/index.c > >>> [0]PETSC ERROR: MatLUFactorSymbolic_SeqAIJ() line 276 in > >>> src/mat/impls/aij/seq/aijfact.c > >>> [0]PETSC ERROR: MatLUFactorSymbolic() line 2178 in > >>> src/mat/interface/matrix.c > >>> [localhost.localdomain:24730] MPI_ABORT invoked on rank 0 in > >>> communicator MPI_COMM_WORLD with errorcode 62 > >>> > >>> Here is the portion of the offending code (I think): > >>> > >>> call MatAssemblyBegin(A,MAT_FINAL_ASSEMBLY,error);CHKERRQ(error) > >>> call MatAssemblyEnd(A,MAT_FINAL_ASSEMBLY,error);CHKERRQ(error) > >>> call ISCreateGeneral(PETSC_COMM_WORLD,order,columnIndices, > >>> indexSet,error);CHKERRQ(error) > >>> call MatFactorInfoInitialize(info,error);CHKERRQ(error) > >>> call > >>> MatLUFactorSymbolic(A,indexSet,indexSet,info,factorMat,error);CHKERRQ(error) > >>> > >>> order - number of rows/columns in square matrix A > >>> columnIndices - 1D array containing values 0-->order-1 > >>> > >>> All types are set correctly as far as I can tell. > >>> > >>> Can anyone suggest why the execution falls over during the > >>> MatLUFactorSymbolic() call? > >>> > >>> Thanks again, > >>> > >>> Tim. > >>> > >>> -- > >>> Dr. Timothy Stitt > >>> HPC Application Consultant - ICHEC (www.ichec.ie) > >>> > >>> Dublin Institute for Advanced Studies > >>> 5 Merrion Square - Dublin 2 - Ireland > >>> > >>> +353-1-6621333 (tel) / +353-1-6621477 (fax) > >>> > >>> > >>> > >> > >> > > > > > > > -- > Dr. Timothy Stitt > HPC Application Consultant - ICHEC (www.ichec.ie) > > Dublin Institute for Advanced Studies > 5 Merrion Square - Dublin 2 - Ireland > > +353-1-6621333 (tel) / +353-1-6621477 (fax) > > -- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. -- Norbert Wiener From bsmith at mcs.anl.gov Thu Oct 25 14:26:54 2007 From: bsmith at mcs.anl.gov (Barry Smith) Date: Thu, 25 Oct 2007 14:26:54 -0500 (CDT) Subject: Invalid argument! - not a permutation! In-Reply-To: <4720ED12.4040708@ichec.ie> References: <4720E65C.9070504@ichec.ie> <4720ED12.4040708@ichec.ie> Message-ID: On Thu, 25 Oct 2007, Tim Stitt wrote: > OK...that is useful thanks. > > If you don't mind just one more question...I notice that the > MatLUFactorSymbolic() routine now executes with one processor but complains > when I run > 2 processors. The actual error message is: > > [1]PETSC ERROR: No support for this operation for this object type! > [1]PETSC ERROR: Matrix type mpiaij symbolic LU! > > Is this because the default LU factorisation routine isn't parallel... Yes > I just > assumed parallel superlu would be picked up automatically (it is compiled into > my PETSc build) ? No, in the current model you have to set the matrix type to MATSUPERLU_DIST to get it. Or MATAIJMUMPS for mumps etc. Hong has suggested an alternative (better?) model, but ..... Barry > > Tim. > > Barry Smith wrote: > > I have added this information to the error message that gets printed. > > > > Barry > > > > > > On Thu, 25 Oct 2007, Matthew Knepley wrote: > > > > > > > I think you need to tell use that the IS is a permutation: > > > > > > http://www-unix.mcs.anl.gov/petsc/petsc-as/snapshots/petsc-current/docs/manualpages/IS/ISSetPermutation.html > > > > > > Matt > > > > > > On 10/25/07, Tim Stitt wrote: > > > > > > > Hi all, > > > > > > > > Following on from some help I got a few days ago with respect to > > > > MatLUFactorSymbolic() usage in Fortran...I think I now have all the > > > > parts in place but keep getting the following runtime error upon > > > > execution: > > > > > > > > [0]PETSC ERROR: --------------------- Error Message > > > > ------------------------------------ > > > > [0]PETSC ERROR: Invalid argument! > > > > [0]PETSC ERROR: not a permutation! > > > > ... > > > > [0]PETSC ERROR: > > > > ------------------------------------------------------------------------ > > > > [0]PETSC ERROR: ISInvertPermutation() line 183 in > > > > src/vec/is/interface/index.c > > > > [0]PETSC ERROR: MatLUFactorSymbolic_SeqAIJ() line 276 in > > > > src/mat/impls/aij/seq/aijfact.c > > > > [0]PETSC ERROR: MatLUFactorSymbolic() line 2178 in > > > > src/mat/interface/matrix.c > > > > [localhost.localdomain:24730] MPI_ABORT invoked on rank 0 in > > > > communicator MPI_COMM_WORLD with errorcode 62 > > > > > > > > Here is the portion of the offending code (I think): > > > > > > > > call MatAssemblyBegin(A,MAT_FINAL_ASSEMBLY,error);CHKERRQ(error) > > > > call MatAssemblyEnd(A,MAT_FINAL_ASSEMBLY,error);CHKERRQ(error) > > > > call ISCreateGeneral(PETSC_COMM_WORLD,order,columnIndices, > > > > indexSet,error);CHKERRQ(error) > > > > call MatFactorInfoInitialize(info,error);CHKERRQ(error) > > > > call > > > > MatLUFactorSymbolic(A,indexSet,indexSet,info,factorMat,error);CHKERRQ(error) > > > > > > > > order - number of rows/columns in square matrix A > > > > columnIndices - 1D array containing values 0-->order-1 > > > > > > > > All types are set correctly as far as I can tell. > > > > > > > > Can anyone suggest why the execution falls over during the > > > > MatLUFactorSymbolic() call? > > > > > > > > Thanks again, > > > > > > > > Tim. > > > > > > > > -- > > > > Dr. Timothy Stitt > > > > HPC Application Consultant - ICHEC (www.ichec.ie) > > > > > > > > Dublin Institute for Advanced Studies > > > > 5 Merrion Square - Dublin 2 - Ireland > > > > > > > > +353-1-6621333 (tel) / +353-1-6621477 (fax) > > > > > > > > > > > > > > > > > > > > > > > > > From keita at cray.com Thu Oct 25 15:19:45 2007 From: keita at cray.com (Keita Teranishi) Date: Thu, 25 Oct 2007 15:19:45 -0500 Subject: LCD method in KSP Message-ID: <925346A443D4E340BEB20248BAFCDBDF02CE21DB@CFEVS1-IP.americas.cray.com> Hi, I am in trouble to find any document on a relatively new method called LCD in KSP even though it listed in the KSP_TYPE document. I am wondering whether LCD is fully supported by the current version of PETSc, Thanks, ================================ Keita Teranishi Math Software Group Cray Inc. keita at cray.com ================================ -------------- next part -------------- An HTML attachment was scrubbed... URL: From bsmith at mcs.anl.gov Thu Oct 25 15:56:54 2007 From: bsmith at mcs.anl.gov (Barry Smith) Date: Thu, 25 Oct 2007 15:56:54 -0500 (CDT) Subject: LCD method in KSP In-Reply-To: <925346A443D4E340BEB20248BAFCDBDF02CE21DB@CFEVS1-IP.americas.cray.com> References: <925346A443D4E340BEB20248BAFCDBDF02CE21DB@CFEVS1-IP.americas.cray.com> Message-ID: Oh, I see it is missing a manual page. I have just written one. Here it is /*MC KSPLCD - Implements the LCD (left conjugate direction) method in PETSc. Options Database Keys: + -ksp_lcd_restart - number of vectors conjudate - -ksp_lcd_haptol - tolerance for exact convergence (happing ending) Level: beginner References: - J.Y. Yuan, G.H.Golub, R.J. Plemmons, and W.A.G. Cecilio. Semiconjugate direction methods for real positive definite system. BIT Numerical Mathematics, 44(1):189-207,2004. - Y. Dai and J.Y. Yuan. Study on semi-conjugate direction methods for non-symmetric systems. International Journal for Numerical Methods in Engineering, 60:1383-1399,2004. - L. Catabriga, A.L.G.A. Coutinho, and L.P.Franca. Evaluating the LCD algorithm for solving linear systems of equations arising from implicit SUPG formulation of compressible flows. International Journal for Numerical Methods in Engineering, 60:1513-1534,2004 - L. Catabriga, A. M. P. Valli, B. Z. Melotti, L. M. Pessoa, A. L. G. A. Coutinho, Performance of LCD iterative method in the finite element and finite difference solution of convection-diffusion equations, Communications in Numerical Methods in Engineering, (Early View). Contributed by: Lucia Catabriga .seealso: KSPCreate(), KSPSetType(), KSPType (for list of available types), KSP, KSPCGSetType(), KSPLCDSetRestart(), KSPLCDSetHapTol() M*/ I know nothing beyond this. Barry On Thu, 25 Oct 2007, Keita Teranishi wrote: > Hi, > > > > I am in trouble to find any document on a relatively new method called LCD in KSP even though it listed in the KSP_TYPE document. I am wondering whether LCD is fully supported by the current version of PETSc, > > > > Thanks, > > ================================ > Keita Teranishi > Math Software Group > Cray Inc. > keita at cray.com > ================================ > > > > From hzhang at mcs.anl.gov Thu Oct 25 15:57:29 2007 From: hzhang at mcs.anl.gov (Hong Zhang) Date: Thu, 25 Oct 2007 15:57:29 -0500 (CDT) Subject: Invalid argument! - not a permutation! In-Reply-To: References: <4720E65C.9070504@ichec.ie> <4720ED12.4040708@ichec.ie> Message-ID: >> >> Is this because the default LU factorisation routine isn't parallel... > > Yes > >> I just >> assumed parallel superlu would be picked up automatically (it is compiled into >> my PETSc build) ? > > No, in the current model you have to set the matrix type to MATSUPERLU_DIST > to get it. Or MATAIJMUMPS for mumps etc. > > Hong has suggested an alternative (better?) model, but ..... Petsc only provides sequential LU factorisation. To run parallel LU, you must use one of the external packages, SUPERLU_DIS, MUMPS or spooles. Simply install petsc with these packages, then run your petsc code with the option '-mat_type superlu_dist' '-mat_type aijmumps' or '-mat_type aijspooles' or set the matrix type 'MATSUPERLU_DIST' in your code. Note: all these packages support sequential implementation. MUMPS sequential LU solver usually is the fastest. As matrix size increases, SUPERLU might catch up. Let us know if you have trouble using them, Hong > > Barry > >> >> Tim. >> >> Barry Smith wrote: >>> I have added this information to the error message that gets printed. >>> >>> Barry >>> >>> >>> On Thu, 25 Oct 2007, Matthew Knepley wrote: >>> >>> >>>> I think you need to tell use that the IS is a permutation: >>>> >>>> http://www-unix.mcs.anl.gov/petsc/petsc-as/snapshots/petsc-current/docs/manualpages/IS/ISSetPermutation.html >>>> >>>> Matt >>>> >>>> On 10/25/07, Tim Stitt wrote: >>>> >>>>> Hi all, >>>>> >>>>> Following on from some help I got a few days ago with respect to >>>>> MatLUFactorSymbolic() usage in Fortran...I think I now have all the >>>>> parts in place but keep getting the following runtime error upon >>>>> execution: >>>>> >>>>> [0]PETSC ERROR: --------------------- Error Message >>>>> ------------------------------------ >>>>> [0]PETSC ERROR: Invalid argument! >>>>> [0]PETSC ERROR: not a permutation! >>>>> ... >>>>> [0]PETSC ERROR: >>>>> ------------------------------------------------------------------------ >>>>> [0]PETSC ERROR: ISInvertPermutation() line 183 in >>>>> src/vec/is/interface/index.c >>>>> [0]PETSC ERROR: MatLUFactorSymbolic_SeqAIJ() line 276 in >>>>> src/mat/impls/aij/seq/aijfact.c >>>>> [0]PETSC ERROR: MatLUFactorSymbolic() line 2178 in >>>>> src/mat/interface/matrix.c >>>>> [localhost.localdomain:24730] MPI_ABORT invoked on rank 0 in >>>>> communicator MPI_COMM_WORLD with errorcode 62 >>>>> >>>>> Here is the portion of the offending code (I think): >>>>> >>>>> call MatAssemblyBegin(A,MAT_FINAL_ASSEMBLY,error);CHKERRQ(error) >>>>> call MatAssemblyEnd(A,MAT_FINAL_ASSEMBLY,error);CHKERRQ(error) >>>>> call ISCreateGeneral(PETSC_COMM_WORLD,order,columnIndices, >>>>> indexSet,error);CHKERRQ(error) >>>>> call MatFactorInfoInitialize(info,error);CHKERRQ(error) >>>>> call >>>>> MatLUFactorSymbolic(A,indexSet,indexSet,info,factorMat,error);CHKERRQ(error) >>>>> >>>>> order - number of rows/columns in square matrix A >>>>> columnIndices - 1D array containing values 0-->order-1 >>>>> >>>>> All types are set correctly as far as I can tell. >>>>> >>>>> Can anyone suggest why the execution falls over during the >>>>> MatLUFactorSymbolic() call? >>>>> >>>>> Thanks again, >>>>> >>>>> Tim. >>>>> >>>>> -- >>>>> Dr. Timothy Stitt >>>>> HPC Application Consultant - ICHEC (www.ichec.ie) >>>>> >>>>> Dublin Institute for Advanced Studies >>>>> 5 Merrion Square - Dublin 2 - Ireland >>>>> >>>>> +353-1-6621333 (tel) / +353-1-6621477 (fax) >>>>> >>>>> >>>>> >>>> >>>> >>> >>> >> >> >> > > From z.sheng at ewi.tudelft.nl Fri Oct 26 08:57:20 2007 From: z.sheng at ewi.tudelft.nl (Zhifeng Sheng) Date: Fri, 26 Oct 2007 15:57:20 +0200 Subject: A^TA + B? Message-ID: <4721F240.6080709@ewi.tudelft.nl> Dear all I would like to implement B = A^TA + B. And the transpose multiplication can be done with the following function, PetscErrorCode PETSCMAT_DLLEXPORT MatMatMultTranspose(Mat A,Mat B,MatReuse scall,PetscReal fill,Mat *C) since I would like to save some memory, I don't want to create an intermediate matrix. In the above function, the discription of the argument "scall" is not clear, I know that MAT_INITIAL_MATRIX will create a new empty matrix and the result will be stored. what about MAT_REUSE_MATRIX ? does it mean that the C matrix will be added with the result? Thank you Best regards Zhifeng Sheng From hzhang at mcs.anl.gov Fri Oct 26 09:45:45 2007 From: hzhang at mcs.anl.gov (Hong Zhang) Date: Fri, 26 Oct 2007 09:45:45 -0500 (CDT) Subject: A^TA + B? In-Reply-To: <4721F240.6080709@ewi.tudelft.nl> References: <4721F240.6080709@ewi.tudelft.nl> Message-ID: Zhifeng, > > I would like to implement B = A^TA + B. > > And the transpose multiplication can be done with the following function, > > PetscErrorCode PETSCMAT_DLLEXPORT MatMatMultTranspose(Mat A,Mat B,MatReuse > scall,PetscReal fill,Mat *C) > since I would like to save some memory, I don't want to create an > intermediate matrix. In the above function, the discription of > > the argument "scall" is not clear, I know that MAT_INITIAL_MATRIX will create > a new empty matrix and the result will be stored. Yes. > > what about MAT_REUSE_MATRIX ? does it mean that the C matrix will be added > with the result? When the product matrix C=A^T*B is previously computed, and you want repeating C=A^T*B, in which A and B maintain the same non-zero sparse pattern but with different numerical values, using MAT_REUSE_MATRIX will skip the symbolic computation of C=A^T*B and reuse the exiting matrix data structure and memory of C. Why do you need B = A^TA + B? If your matrices are dense, you should use lapack or scalapack. Sparse matrix product should be avoided because the product matrix C usually is much denser than A and B, and C=A^T*B cannot be implemented efficiently in general. We provide MatMatMultTranspose() mainly for multigrid computation in which the matrices have special data structure. Best, Hong From z.sheng at ewi.tudelft.nl Fri Oct 26 11:07:01 2007 From: z.sheng at ewi.tudelft.nl (Zhifeng Sheng) Date: Fri, 26 Oct 2007 18:07:01 +0200 Subject: A^TA + B? In-Reply-To: References: <4721F240.6080709@ewi.tudelft.nl> Message-ID: <472210A5.2080500@ewi.tudelft.nl> Hong Zhang wrote: > > Zhifeng, > >> >> I would like to implement B = A^TA + B. >> >> And the transpose multiplication can be done with the following >> function, >> >> PetscErrorCode PETSCMAT_DLLEXPORT MatMatMultTranspose(Mat A,Mat >> B,MatReuse scall,PetscReal fill,Mat *C) since I would like to save >> some memory, I don't want to create an intermediate matrix. In the >> above function, the discription of >> >> the argument "scall" is not clear, I know that MAT_INITIAL_MATRIX >> will create a new empty matrix and the result will be stored. > > > Yes. > >> >> what about MAT_REUSE_MATRIX ? does it mean that the C matrix will be >> added with the result? > > > When the product matrix C=A^T*B is previously computed, > and you want repeating C=A^T*B, in which A and B maintain > the same non-zero sparse pattern but with different numerical values, > using > MAT_REUSE_MATRIX will skip the symbolic computation of C=A^T*B > and reuse the exiting matrix data structure and memory of C. > > Why do you need B = A^TA + B? > If your matrices are dense, you should use lapack or scalapack. > Sparse matrix product should be avoided because > the product matrix C usually is much denser than A and B, > and C=A^T*B cannot be implemented efficiently in general. > We provide MatMatMultTranspose() mainly for multigrid computation > in which the matrices have special data structure. > > Best, > > Hong > Thanks for your reply. I need to implement a least-squares method, in which A^T*A will be computed, however, since I use finite elements, local matrices will be computed and then added to the global matrix. The A matrix is very sparse, e.g 9 non-zeros in a row. Actually, in this case, I only need to do b = a^T*a + b on local matrices, which are very small e.g (12 by 12). The elementary matrices will be assembled in the end. Should I use lapack in combination with Petsc? Best regards Zhifeng From dalcinl at gmail.com Fri Oct 26 12:55:12 2007 From: dalcinl at gmail.com (Lisandro Dalcin) Date: Fri, 26 Oct 2007 14:55:12 -0300 Subject: A^TA + B? In-Reply-To: <472210A5.2080500@ewi.tudelft.nl> References: <4721F240.6080709@ewi.tudelft.nl> <472210A5.2080500@ewi.tudelft.nl> Message-ID: Perhaps I missunderstood you, but... Why don't you assemble your global B matrix from the elemental matrices a^Ta+b ? An then you should definitelly use as sparse AIJ matrix for this for this and solve it with an iterative method, but not a dense matrix with lapack. On 10/26/07, Zhifeng Sheng wrote: > Hong Zhang wrote: > Actually, in this case, I only need to do b = a^T*a + b on local > matrices, which are very small e.g (12 by 12). The elementary matrices > will be assembled in the end. Should I use lapack in combination with Petsc? > -- Lisandro Dalc?n --------------- Centro Internacional de M?todos Computacionales en Ingenier?a (CIMEC) Instituto de Desarrollo Tecnol?gico para la Industria Qu?mica (INTEC) Consejo Nacional de Investigaciones Cient?ficas y T?cnicas (CONICET) PTLC - G?emes 3450, (3000) Santa Fe, Argentina Tel/Fax: +54-(0)342-451.1594 From jwicks at cs.brown.edu Fri Oct 26 13:56:15 2007 From: jwicks at cs.brown.edu (John R. Wicks) Date: Fri, 26 Oct 2007 14:56:15 -0400 Subject: Slow assembly Message-ID: <000e01c81801$e3c17700$6b01a8c0@jwickslptp> I am working on computing PageRank for a web scale graph which uses a square matrix which is 1.2x10^8 dimensional with about 10^9 entries. I have partitioned the matrix for 32 processors myself into my own ascii format, and I know the memory allocation, so I: 1) create the matrix with "A = MatCreateMPIAIJ(*n, *n, *N, *N, 0, nnz, 0, onnz)", 2) load the entries by repeatedly calling "MatSetValues(A,1,&row,links,cols,vals,INSERT_VALUES)", and 3) call MatAssemblyBegin/End. Steps 1 and 2 complete in a couple minutes, but step 3 is taking several hours. What is going on? Is there a way to speed up matrix assembly? Thanks, John From balay at mcs.anl.gov Fri Oct 26 14:04:03 2007 From: balay at mcs.anl.gov (Satish Balay) Date: Fri, 26 Oct 2007 14:04:03 -0500 (CDT) Subject: Slow assembly In-Reply-To: <000e01c81801$e3c17700$6b01a8c0@jwickslptp> References: <000e01c81801$e3c17700$6b01a8c0@jwickslptp> Message-ID: On Fri, 26 Oct 2007, John R. Wicks wrote: > I am working on computing PageRank for a web scale graph which uses a square > matrix which is 1.2x10^8 dimensional with about 10^9 entries. I have > partitioned the matrix for 32 processors myself into my own ascii format, > and I know the memory allocation, so I: > > 1) create the matrix with "A = MatCreateMPIAIJ(*n, *n, *N, *N, 0, nnz, 0, > onnz)", > 2) load the entries by repeatedly calling > "MatSetValues(A,1,&row,links,cols,vals,INSERT_VALUES)", and > > 3) call MatAssemblyBegin/End. > > Steps 1 and 2 complete in a couple minutes, but step 3 is taking several > hours. What is going on? Is there a way to speed up matrix assembly? Are you makeing sure that you call MatGetOwnershipRange() - and calling MatSetValues() for mostly local rows only? Also can you confirm that multiple processes [for eg: proc-0 and proc-1 etc..] are not setting the same value [i.e both of them calling MatSetValues(row=0,col=0)] Satish From jwicks at cs.brown.edu Fri Oct 26 14:49:04 2007 From: jwicks at cs.brown.edu (John R. Wicks) Date: Fri, 26 Oct 2007 15:49:04 -0400 Subject: Slow assembly In-Reply-To: Message-ID: <000f01c81809$442d7060$6b01a8c0@jwickslptp> I have confirmed that I am calling MatSetValues() for local rows only and am only setting each value exactly once. Because of how the matrix was partitioned for another non-Petsc program, each partition is partitioned (by columns) into 32 blocks (corresponding to the row partitions). I enter the data for each block one row at a time, i.e., for any one SetValues call, the entries are sorted by increasing column index. Does that mean I can use MatrixSetOption(A,MAT_COLUMNS_SORTED). Should that help? P.S.: I tried it, and it still seems to be taking quite a long time. > -----Original Message----- > From: owner-petsc-users at mcs.anl.gov > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Satish Balay > Sent: Friday, October 26, 2007 3:04 PM > To: petsc-users at mcs.anl.gov > Subject: Re: Slow assembly > > > On Fri, 26 Oct 2007, John R. Wicks wrote: > > > I am working on computing PageRank for a web scale graph > which uses a > > square matrix which is 1.2x10^8 dimensional with about 10^9 > entries. > > I have partitioned the matrix for 32 processors myself into my own > > ascii format, and I know the memory allocation, so I: > > > > 1) create the matrix with "A = MatCreateMPIAIJ(*n, *n, *N, > *N, 0, nnz, > > 0, onnz)", > > 2) load the entries by repeatedly calling > > "MatSetValues(A,1,&row,links,cols,vals,INSERT_VALUES)", and > > > > 3) call MatAssemblyBegin/End. > > > > Steps 1 and 2 complete in a couple minutes, but step 3 is taking > > several hours. What is going on? Is there a way to speed > up matrix > > assembly? > > Are you makeing sure that you call MatGetOwnershipRange() - > and calling MatSetValues() for mostly local rows only? > > Also can you confirm that multiple processes [for eg: proc-0 > and proc-1 etc..] are not setting the same value [i.e both > of them calling MatSetValues(row=0,col=0)] > > Satish > From balay at mcs.anl.gov Fri Oct 26 14:58:04 2007 From: balay at mcs.anl.gov (Satish Balay) Date: Fri, 26 Oct 2007 14:58:04 -0500 (CDT) Subject: Slow assembly In-Reply-To: <000f01c81809$442d7060$6b01a8c0@jwickslptp> References: <000f01c81809$442d7060$6b01a8c0@jwickslptp> Message-ID: If matassembly takes a long time - it usually means that lots of values are set on the wrong processor [they get communicated to the correct processor during assembly]. But several hours is a too long for this operation [so I don't understand whats hapenning here] Can you run the code with the option '-info' - and then send us this log at petsc-maint at mcs.anl.gov Satish On Fri, 26 Oct 2007, John R. Wicks wrote: > I have confirmed that I am calling MatSetValues() for local rows only and am > only setting each value exactly once. > > Because of how the matrix was partitioned for another non-Petsc program, > each partition is partitioned (by columns) into 32 blocks (corresponding to > the row partitions). I enter the data for each block one row at a time, > i.e., for any one SetValues call, the entries are sorted by increasing > column index. Does that mean I can use > MatrixSetOption(A,MAT_COLUMNS_SORTED). Should that help? > > P.S.: I tried it, and it still seems to be taking quite a long time. > > > -----Original Message----- > > From: owner-petsc-users at mcs.anl.gov > > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Satish Balay > > Sent: Friday, October 26, 2007 3:04 PM > > To: petsc-users at mcs.anl.gov > > Subject: Re: Slow assembly > > > > > > On Fri, 26 Oct 2007, John R. Wicks wrote: > > > > > I am working on computing PageRank for a web scale graph > > which uses a > > > square matrix which is 1.2x10^8 dimensional with about 10^9 > > entries. > > > I have partitioned the matrix for 32 processors myself into my own > > > ascii format, and I know the memory allocation, so I: > > > > > > 1) create the matrix with "A = MatCreateMPIAIJ(*n, *n, *N, > > *N, 0, nnz, > > > 0, onnz)", > > > 2) load the entries by repeatedly calling > > > "MatSetValues(A,1,&row,links,cols,vals,INSERT_VALUES)", and > > > > > > 3) call MatAssemblyBegin/End. > > > > > > Steps 1 and 2 complete in a couple minutes, but step 3 is taking > > > several hours. What is going on? Is there a way to speed > > up matrix > > > assembly? > > > > Are you makeing sure that you call MatGetOwnershipRange() - > > and calling MatSetValues() for mostly local rows only? > > > > Also can you confirm that multiple processes [for eg: proc-0 > > and proc-1 etc..] are not setting the same value [i.e both > > of them calling MatSetValues(row=0,col=0)] > > > > Satish > > > > From bsmith at mcs.anl.gov Fri Oct 26 18:53:51 2007 From: bsmith at mcs.anl.gov (Barry Smith) Date: Fri, 26 Oct 2007 18:53:51 -0500 (CDT) Subject: Slow assembly In-Reply-To: <000f01c81809$442d7060$6b01a8c0@jwickslptp> References: <000f01c81809$442d7060$6b01a8c0@jwickslptp> Message-ID: The sorting does not matter. Under normal conditions the MatAssembly should take a fraction of a second. The only cause that we know that slows it down to the extreme you have is that it is sending a huge amount of data across processes (the -info option Satish suggested will tell us if that is true). Are you only call MatAssemblyBegin/End() once? You should, don't call it multiple times. The sorting is not important (in fact it takes advantage of it automatically and does not need to be set). Barry On Fri, 26 Oct 2007, John R. Wicks wrote: > I have confirmed that I am calling MatSetValues() for local rows only and am > only setting each value exactly once. > > Because of how the matrix was partitioned for another non-Petsc program, > each partition is partitioned (by columns) into 32 blocks (corresponding to > the row partitions). I enter the data for each block one row at a time, > i.e., for any one SetValues call, the entries are sorted by increasing > column index. Does that mean I can use > MatrixSetOption(A,MAT_COLUMNS_SORTED). Should that help? > > P.S.: I tried it, and it still seems to be taking quite a long time. > > > -----Original Message----- > > From: owner-petsc-users at mcs.anl.gov > > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Satish Balay > > Sent: Friday, October 26, 2007 3:04 PM > > To: petsc-users at mcs.anl.gov > > Subject: Re: Slow assembly > > > > > > On Fri, 26 Oct 2007, John R. Wicks wrote: > > > > > I am working on computing PageRank for a web scale graph > > which uses a > > > square matrix which is 1.2x10^8 dimensional with about 10^9 > > entries. > > > I have partitioned the matrix for 32 processors myself into my own > > > ascii format, and I know the memory allocation, so I: > > > > > > 1) create the matrix with "A = MatCreateMPIAIJ(*n, *n, *N, > > *N, 0, nnz, > > > 0, onnz)", > > > 2) load the entries by repeatedly calling > > > "MatSetValues(A,1,&row,links,cols,vals,INSERT_VALUES)", and > > > > > > 3) call MatAssemblyBegin/End. > > > > > > Steps 1 and 2 complete in a couple minutes, but step 3 is taking > > > several hours. What is going on? Is there a way to speed > > up matrix > > > assembly? > > > > Are you makeing sure that you call MatGetOwnershipRange() - > > and calling MatSetValues() for mostly local rows only? > > > > Also can you confirm that multiple processes [for eg: proc-0 > > and proc-1 etc..] are not setting the same value [i.e both > > of them calling MatSetValues(row=0,col=0)] > > > > Satish > > > > From jwicks at cs.brown.edu Sun Oct 28 15:21:16 2007 From: jwicks at cs.brown.edu (John R. Wicks) Date: Sun, 28 Oct 2007 16:21:16 -0400 Subject: Slow assembly In-Reply-To: Message-ID: <000001c819a0$19f09c80$6b01a8c0@jwickslptp> I ran on a smaller example with the -info switch set. I preallocated all my matrices (there are two of them, one sequential and one distributed) so that there should be no allocs needed, but I notice it reports: something like: [3] MatAssemblyEnd_SeqAIJ(): Matrix size: 11 X 11; storage space: 10 unneeded,11 used [3] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 [3] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 [1] MatAssemblyEnd_SeqAIJ(): Matrix size: 10 X 10; storage space: 13 unneeded,10 used [1] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 [1] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 [2] MatAssemblyEnd_SeqAIJ(): Matrix size: 11 X 11; storage space: 22 unneeded,11 used [2] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 [2] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 [5] MatAssemblyEnd_SeqAIJ(): Matrix size: 10 X 10; storage space: 9 unneeded,10 used [5] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 [5] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 [4] MatAssemblyEnd_SeqAIJ(): Matrix size: 9 X 9; storage space: 13 unneeded,9 used [4] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 [4] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 [0] MatAssemblyEnd_SeqAIJ(): Matrix size: 11 X 11; storage space: 19 unneeded,11 used [0] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 [0] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 [6] MatAssemblyEnd_SeqAIJ(): Matrix size: 10 X 10; storage space: 17 unneeded,10 used [6] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 [6] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 [2] MatAssemblyEnd_SeqAIJ(): Matrix size: 11 X 11; storage space: 115 unneeded,31 used [2] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 9 [2] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 5 [5] MatAssemblyEnd_SeqAIJ(): Matrix size: 10 X 10; storage space: 67 unneeded,18 used [5] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 5 [5] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 3 [0] MatAssemblyEnd_SeqAIJ(): Matrix size: 11 X 11; storage space: 88 unneeded,28 used [0] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 7 [0] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 5 I can't figure why it should allocate anything, b/c I've precomputed the number of entries in each row. > -----Original Message----- > From: owner-petsc-users at mcs.anl.gov > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Barry Smith > Sent: Friday, October 26, 2007 7:54 PM > To: petsc-users at mcs.anl.gov > Subject: RE: Slow assembly > > > > The sorting does not matter. > > Under normal conditions the MatAssembly should take a > fraction of a second. The only cause that we know that slows > it down to the extreme you have is that it is sending a huge > amount of data across processes (the -info option Satish > suggested will tell us if that is true). > > Are you only call MatAssemblyBegin/End() once? You should, > don't call it > multiple times. > > The sorting is not important (in fact it takes advantage of > it automatically and does not need to be set). > > Barry > > > On Fri, 26 Oct 2007, John R. Wicks wrote: > > > I have confirmed that I am calling MatSetValues() for local > rows only > > and am only setting each value exactly once. > > > > Because of how the matrix was partitioned for another non-Petsc > > program, each partition is partitioned (by columns) into 32 blocks > > (corresponding to the row partitions). I enter the data for each > > block one row at a time, i.e., for any one SetValues call, > the entries > > are sorted by increasing column index. Does that mean I can use > > MatrixSetOption(A,MAT_COLUMNS_SORTED). Should that help? > > > > P.S.: I tried it, and it still seems to be taking quite a long time. > > > > > -----Original Message----- > > > From: owner-petsc-users at mcs.anl.gov > > > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Satish Balay > > > Sent: Friday, October 26, 2007 3:04 PM > > > To: petsc-users at mcs.anl.gov > > > Subject: Re: Slow assembly > > > > > > > > > On Fri, 26 Oct 2007, John R. Wicks wrote: > > > > > > > I am working on computing PageRank for a web scale graph > > > which uses a > > > > square matrix which is 1.2x10^8 dimensional with about 10^9 > > > entries. > > > > I have partitioned the matrix for 32 processors myself > into my own > > > > ascii format, and I know the memory allocation, so I: > > > > > > > > 1) create the matrix with "A = MatCreateMPIAIJ(*n, *n, *N, > > > *N, 0, nnz, > > > > 0, onnz)", > > > > 2) load the entries by repeatedly calling > > > > "MatSetValues(A,1,&row,links,cols,vals,INSERT_VALUES)", and > > > > > > > > 3) call MatAssemblyBegin/End. > > > > > > > > Steps 1 and 2 complete in a couple minutes, but step 3 is taking > > > > several hours. What is going on? Is there a way to speed > > > up matrix > > > > assembly? > > > > > > Are you makeing sure that you call MatGetOwnershipRange() - > > > and calling MatSetValues() for mostly local rows only? > > > > > > Also can you confirm that multiple processes [for eg: proc-0 > > > and proc-1 etc..] are not setting the same value [i.e both > > > of them calling MatSetValues(row=0,col=0)] > > > > > > Satish > > > > > > > > From knepley at gmail.com Sun Oct 28 15:28:37 2007 From: knepley at gmail.com (Matthew Knepley) Date: Sun, 28 Oct 2007 15:28:37 -0500 Subject: Slow assembly In-Reply-To: <000001c819a0$19f09c80$6b01a8c0@jwickslptp> References: <000001c819a0$19f09c80$6b01a8c0@jwickslptp> Message-ID: Perhaps this is a miscalculation for parallel allocation, meaning you allocate the correct number of values, but do not divide them correctly between the diagonal and offdiagonal parts. Matt On 10/28/07, John R. Wicks wrote: > I ran on a smaller example with the -info switch set. I preallocated all my > matrices (there are two of them, one sequential and one distributed) so that > there should be no allocs needed, but I notice it reports: something like: > [3] MatAssemblyEnd_SeqAIJ(): Matrix size: 11 X 11; storage space: 10 > unneeded,11 used > [3] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 > [3] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 > [1] MatAssemblyEnd_SeqAIJ(): Matrix size: 10 X 10; storage space: 13 > unneeded,10 used > [1] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 > [1] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 > [2] MatAssemblyEnd_SeqAIJ(): Matrix size: 11 X 11; storage space: 22 > unneeded,11 used > [2] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 > [2] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 > [5] MatAssemblyEnd_SeqAIJ(): Matrix size: 10 X 10; storage space: 9 > unneeded,10 used > [5] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 > [5] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 > [4] MatAssemblyEnd_SeqAIJ(): Matrix size: 9 X 9; storage space: 13 > unneeded,9 used > [4] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 > [4] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 > [0] MatAssemblyEnd_SeqAIJ(): Matrix size: 11 X 11; storage space: 19 > unneeded,11 used > [0] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 > [0] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 > [6] MatAssemblyEnd_SeqAIJ(): Matrix size: 10 X 10; storage space: 17 > unneeded,10 used > [6] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 > [6] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 > [2] MatAssemblyEnd_SeqAIJ(): Matrix size: 11 X 11; storage space: 115 > unneeded,31 used > [2] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 9 > [2] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 5 > [5] MatAssemblyEnd_SeqAIJ(): Matrix size: 10 X 10; storage space: 67 > unneeded,18 used > [5] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 5 > [5] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 3 > [0] MatAssemblyEnd_SeqAIJ(): Matrix size: 11 X 11; storage space: 88 > unneeded,28 used > [0] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 7 > [0] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 5 > > I can't figure why it should allocate anything, b/c I've precomputed the > number of entries in each row. > > > -----Original Message----- > > From: owner-petsc-users at mcs.anl.gov > > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Barry Smith > > Sent: Friday, October 26, 2007 7:54 PM > > To: petsc-users at mcs.anl.gov > > Subject: RE: Slow assembly > > > > > > > > The sorting does not matter. > > > > Under normal conditions the MatAssembly should take a > > fraction of a second. The only cause that we know that slows > > it down to the extreme you have is that it is sending a huge > > amount of data across processes (the -info option Satish > > suggested will tell us if that is true). > > > > Are you only call MatAssemblyBegin/End() once? You should, > > don't call it > > multiple times. > > > > The sorting is not important (in fact it takes advantage of > > it automatically and does not need to be set). > > > > Barry > > > > > > On Fri, 26 Oct 2007, John R. Wicks wrote: > > > > > I have confirmed that I am calling MatSetValues() for local > > rows only > > > and am only setting each value exactly once. > > > > > > Because of how the matrix was partitioned for another non-Petsc > > > program, each partition is partitioned (by columns) into 32 blocks > > > (corresponding to the row partitions). I enter the data for each > > > block one row at a time, i.e., for any one SetValues call, > > the entries > > > are sorted by increasing column index. Does that mean I can use > > > MatrixSetOption(A,MAT_COLUMNS_SORTED). Should that help? > > > > > > P.S.: I tried it, and it still seems to be taking quite a long time. > > > > > > > -----Original Message----- > > > > From: owner-petsc-users at mcs.anl.gov > > > > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Satish Balay > > > > Sent: Friday, October 26, 2007 3:04 PM > > > > To: petsc-users at mcs.anl.gov > > > > Subject: Re: Slow assembly > > > > > > > > > > > > On Fri, 26 Oct 2007, John R. Wicks wrote: > > > > > > > > > I am working on computing PageRank for a web scale graph > > > > which uses a > > > > > square matrix which is 1.2x10^8 dimensional with about 10^9 > > > > entries. > > > > > I have partitioned the matrix for 32 processors myself > > into my own > > > > > ascii format, and I know the memory allocation, so I: > > > > > > > > > > 1) create the matrix with "A = MatCreateMPIAIJ(*n, *n, *N, > > > > *N, 0, nnz, > > > > > 0, onnz)", > > > > > 2) load the entries by repeatedly calling > > > > > "MatSetValues(A,1,&row,links,cols,vals,INSERT_VALUES)", and > > > > > > > > > > 3) call MatAssemblyBegin/End. > > > > > > > > > > Steps 1 and 2 complete in a couple minutes, but step 3 is taking > > > > > several hours. What is going on? Is there a way to speed > > > > up matrix > > > > > assembly? > > > > > > > > Are you makeing sure that you call MatGetOwnershipRange() - > > > > and calling MatSetValues() for mostly local rows only? > > > > > > > > Also can you confirm that multiple processes [for eg: proc-0 > > > > and proc-1 etc..] are not setting the same value [i.e both > > > > of them calling MatSetValues(row=0,col=0)] > > > > > > > > Satish > > > > > > > > > > > > > > -- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. -- Norbert Wiener From jwicks at cs.brown.edu Sun Oct 28 22:01:39 2007 From: jwicks at cs.brown.edu (John R. Wicks) Date: Sun, 28 Oct 2007 23:01:39 -0400 Subject: Slow assembly In-Reply-To: Message-ID: <000301c819d8$072d9fc0$6b01a8c0@jwickslptp> What happened was that I used MatDiagonalSet and then used MatSetValues and MatAssemblyEnd, but the log shows that MatAssemblyEnd was called twice; MatDiagonal must force an implicit MatAssemblyBegin/End. By avoiding the call to MatDiagonalSet, everything worked fine (in particular, that shows that my preallocations were correct, as was). > -----Original Message----- > From: owner-petsc-users at mcs.anl.gov > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Matthew Knepley > Sent: Sunday, October 28, 2007 4:29 PM > To: petsc-users at mcs.anl.gov > Subject: Re: Slow assembly > > > Perhaps this is a miscalculation for parallel allocation, > meaning you allocate the correct number of values, but do > not divide them correctly between the diagonal and offdiagonal parts. > > Matt > > On 10/28/07, John R. Wicks wrote: > > I ran on a smaller example with the -info switch set. I > preallocated > > all my matrices (there are two of them, one sequential and one > > distributed) so that there should be no allocs needed, but > I notice it > > reports: something like: [3] MatAssemblyEnd_SeqAIJ(): > Matrix size: 11 > > X 11; storage space: 10 unneeded,11 used [3] > MatAssemblyEnd_SeqAIJ(): > > Number of mallocs during MatSetValues() is 0 [3] > > MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 [1] > > MatAssemblyEnd_SeqAIJ(): Matrix size: 10 X 10; storage space: 13 > > unneeded,10 used [1] MatAssemblyEnd_SeqAIJ(): Number of > mallocs during > > MatSetValues() is 0 [1] MatAssemblyEnd_SeqAIJ(): Maximum > nonzeros in > > any row is 1 [2] MatAssemblyEnd_SeqAIJ(): Matrix size: 11 X 11; > > storage space: 22 unneeded,11 used > > [2] MatAssemblyEnd_SeqAIJ(): Number of mallocs during > MatSetValues() is 0 > > [2] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 > > [5] MatAssemblyEnd_SeqAIJ(): Matrix size: 10 X 10; storage space: 9 > > unneeded,10 used > > [5] MatAssemblyEnd_SeqAIJ(): Number of mallocs during > MatSetValues() is 0 > > [5] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 > > [4] MatAssemblyEnd_SeqAIJ(): Matrix size: 9 X 9; storage space: 13 > > unneeded,9 used > > [4] MatAssemblyEnd_SeqAIJ(): Number of mallocs during > MatSetValues() is 0 > > [4] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 > > [0] MatAssemblyEnd_SeqAIJ(): Matrix size: 11 X 11; storage space: 19 > > unneeded,11 used > > [0] MatAssemblyEnd_SeqAIJ(): Number of mallocs during > MatSetValues() is 0 > > [0] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 > > [6] MatAssemblyEnd_SeqAIJ(): Matrix size: 10 X 10; storage space: 17 > > unneeded,10 used > > [6] MatAssemblyEnd_SeqAIJ(): Number of mallocs during > MatSetValues() is 0 > > [6] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 > > [2] MatAssemblyEnd_SeqAIJ(): Matrix size: 11 X 11; storage > space: 115 > > unneeded,31 used > > [2] MatAssemblyEnd_SeqAIJ(): Number of mallocs during > MatSetValues() is 9 > > [2] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 5 > > [5] MatAssemblyEnd_SeqAIJ(): Matrix size: 10 X 10; storage space: 67 > > unneeded,18 used > > [5] MatAssemblyEnd_SeqAIJ(): Number of mallocs during > MatSetValues() is 5 > > [5] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 3 > > [0] MatAssemblyEnd_SeqAIJ(): Matrix size: 11 X 11; storage space: 88 > > unneeded,28 used > > [0] MatAssemblyEnd_SeqAIJ(): Number of mallocs during > MatSetValues() is 7 > > [0] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 5 > > > > I can't figure why it should allocate anything, b/c I've > precomputed > > the number of entries in each row. > > > > > -----Original Message----- > > > From: owner-petsc-users at mcs.anl.gov > > > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Barry Smith > > > Sent: Friday, October 26, 2007 7:54 PM > > > To: petsc-users at mcs.anl.gov > > > Subject: RE: Slow assembly > > > > > > > > > > > > The sorting does not matter. > > > > > > Under normal conditions the MatAssembly should take a > fraction of > > > a second. The only cause that we know that slows it down to the > > > extreme you have is that it is sending a huge amount of > data across > > > processes (the -info option Satish suggested will tell us > if that is > > > true). > > > > > > Are you only call MatAssemblyBegin/End() once? You > should, don't > > > call it multiple times. > > > > > > The sorting is not important (in fact it takes advantage of it > > > automatically and does not need to be set). > > > > > > Barry > > > > > > > > > On Fri, 26 Oct 2007, John R. Wicks wrote: > > > > > > > I have confirmed that I am calling MatSetValues() for local > > > rows only > > > > and am only setting each value exactly once. > > > > > > > > Because of how the matrix was partitioned for another non-Petsc > > > > program, each partition is partitioned (by columns) > into 32 blocks > > > > (corresponding to the row partitions). I enter the > data for each > > > > block one row at a time, i.e., for any one SetValues call, > > > the entries > > > > are sorted by increasing column index. Does that mean > I can use > > > > MatrixSetOption(A,MAT_COLUMNS_SORTED). Should that help? > > > > > > > > P.S.: I tried it, and it still seems to be taking quite a long > > > > time. > > > > > > > > > -----Original Message----- > > > > > From: owner-petsc-users at mcs.anl.gov > > > > > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of > Satish Balay > > > > > Sent: Friday, October 26, 2007 3:04 PM > > > > > To: petsc-users at mcs.anl.gov > > > > > Subject: Re: Slow assembly > > > > > > > > > > > > > > > On Fri, 26 Oct 2007, John R. Wicks wrote: > > > > > > > > > > > I am working on computing PageRank for a web scale graph > > > > > which uses a > > > > > > square matrix which is 1.2x10^8 dimensional with about 10^9 > > > > > entries. > > > > > > I have partitioned the matrix for 32 processors myself > > > into my own > > > > > > ascii format, and I know the memory allocation, so I: > > > > > > > > > > > > 1) create the matrix with "A = MatCreateMPIAIJ(*n, *n, *N, > > > > > *N, 0, nnz, > > > > > > 0, onnz)", > > > > > > 2) load the entries by repeatedly calling > > > > > > "MatSetValues(A,1,&row,links,cols,vals,INSERT_VALUES)", and > > > > > > > > > > > > 3) call MatAssemblyBegin/End. > > > > > > > > > > > > Steps 1 and 2 complete in a couple minutes, but step 3 is > > > > > > taking several hours. What is going on? Is there a way to > > > > > > speed > > > > > up matrix > > > > > > assembly? > > > > > > > > > > Are you makeing sure that you call > MatGetOwnershipRange() - and > > > > > calling MatSetValues() for mostly local rows only? > > > > > > > > > > Also can you confirm that multiple processes [for eg: > proc-0 and > > > > > proc-1 etc..] are not setting the same value [i.e > both of them > > > > > calling MatSetValues(row=0,col=0)] > > > > > > > > > > Satish > > > > > > > > > > > > > > > > > > > > > > > -- > What most experimenters take for granted before they begin > their experiments is infinitely more interesting than any > results to which their experiments lead. > -- Norbert Wiener > From jwicks at cs.brown.edu Mon Oct 29 05:22:52 2007 From: jwicks at cs.brown.edu (John R. Wicks) Date: Mon, 29 Oct 2007 06:22:52 -0400 Subject: Slow assembly In-Reply-To: Message-ID: <000601c81a15$abbc0300$6b01a8c0@jwickslptp> Although I have the malloc problem fixed, I'm wondering about the number of messages during assembly: [0] MatStashScatterBegin_Private(): No of messages: 0 [0] MatAssemblyBegin_MPIAIJ(): Stash has 0 entries, uses 0 mallocs. [1] MatStashScatterBegin_Private(): No of messages: 1 [1] MatStashScatterBegin_Private(): Mesg_to: 0: size: 272 [1] MatAssemblyBegin_MPIAIJ(): Stash has 33 entries, uses 0 mallocs. [3] MatStashScatterBegin_Private(): No of messages: 3 [3] MatStashScatterBegin_Private(): Mesg_to: 0: size: 360 [3] MatStashScatterBegin_Private(): Mesg_to: 1: size: 328 [3] MatStashScatterBegin_Private(): Mesg_to: 2: size: 360 [3] MatAssemblyBegin_MPIAIJ(): Stash has 128 entries, uses 0 mallocs. [0] MatAssemblyEnd_SeqAIJ(): Matrix size: 11 X 11; storage space: 0 unneeded,11 used [0] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 [0] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 [6] MatStashScatterBegin_Private(): No of messages: 6 [6] MatStashScatterBegin_Private(): Mesg_to: 0: size: 184 [6] MatStashScatterBegin_Private(): Mesg_to: 1: size: 168 [6] MatStashScatterBegin_Private(): Mesg_to: 2: size: 184 [6] MatStashScatterBegin_Private(): Mesg_to: 3: size: 184 [6] MatStashScatterBegin_Private(): Mesg_to: 4: size: 152 [6] MatStashScatterBegin_Private(): Mesg_to: 5: size: 168 [6] MatAssemblyBegin_MPIAIJ(): Stash has 124 entries, uses 0 mallocs. [2] MatStashScatterBegin_Private(): No of messages: 2 [2] MatStashScatterBegin_Private(): Mesg_to: 0: size: 96 [2] MatStashScatterBegin_Private(): Mesg_to: 1: size: 88 [2] MatAssemblyBegin_MPIAIJ(): Stash has 21 entries, uses 0 mallocs. [2] MatAssemblyEnd_SeqAIJ(): Matrix size: 11 X 11; storage space: 0 unneeded,11 used [2] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 [2] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 [5] MatStashScatterBegin_Private(): No of messages: 5 [5] MatStashScatterBegin_Private(): Mesg_to: 0: size: 360 [5] MatStashScatterBegin_Private(): Mesg_to: 1: size: 328 [5] MatStashScatterBegin_Private(): Mesg_to: 2: size: 360 [5] MatStashScatterBegin_Private(): Mesg_to: 3: size: 360 [5] MatStashScatterBegin_Private(): Mesg_to: 4: size: 296 [5] MatAssemblyBegin_MPIAIJ(): Stash has 208 entries, uses 0 mallocs. [5] MatAssemblyEnd_SeqAIJ(): Matrix size: 10 X 10; storage space: 0 unneeded,10 used [5] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 [5] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 [6] MatAssemblyEnd_SeqAIJ(): Matrix size: 10 X 10; storage space: 0 unneeded,10 used [6] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 [6] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 [1] MatAssemblyEnd_SeqAIJ(): Matrix size: 10 X 10; storage space: 0 unneeded,10 used [1] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 [1] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 [4] MatStashScatterBegin_Private(): No of messages: 4 [4] MatStashScatterBegin_Private(): Mesg_to: 0: size: 184 [4] MatStashScatterBegin_Private(): Mesg_to: 1: size: 168 [4] MatStashScatterBegin_Private(): Mesg_to: 2: size: 184 [4] MatStashScatterBegin_Private(): Mesg_to: 3: size: 184 [4] MatAssemblyBegin_MPIAIJ(): Stash has 86 entries, uses 0 mallocs. [3] MatAssemblyEnd_SeqAIJ(): Matrix size: 11 X 11; storage space: 0 unneeded,11 used [3] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 [3] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 [4] MatAssemblyEnd_SeqAIJ(): Matrix size: 9 X 9; storage space: 0 unneeded,9 used [4] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 [4] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 [7] MatStashScatterBegin_Private(): No of messages: 7 [7] MatStashScatterBegin_Private(): Mesg_to: 0: size: 184 [7] MatStashScatterBegin_Private(): Mesg_to: 1: size: 168 [7] MatStashScatterBegin_Private(): Mesg_to: 2: size: 184 [7] MatStashScatterBegin_Private(): Mesg_to: 3: size: 184 [7] MatStashScatterBegin_Private(): Mesg_to: 4: size: 152 [7] MatStashScatterBegin_Private(): Mesg_to: 5: size: 168 [7] MatStashScatter[6] PetscCommDuplicate(): Using internal PETSc communicator 1140850689 -2080374784 [6] PetscCommDuplicate(): returning tag 2147483643 [5] PetscCommDuplicate(): Using internal PETSc communicator 1140850689 -2080374784 [5] PetscCommDuplicate(): returning tag 2147483643 [2] PetscCommDuplicate(): Using internal PETSc communicator 1140850689 -2080374784 [2] PetscCommDuplicate(): returning tag 2147483643 [6] PetscCommDuplicate(): returning tag 2147483643 [5] PetscCommDuplicate(): returning tag 2147483643 [2] PetscCommDuplicate(): returning tag 2147483643 Begin_Private(): Mesg_to: 6: size: 168 [7] MatAssemblyBegin_MPIAIJ(): Stash has 144 entries, uses 0 mallocs. [7] MatAssemblyEnd_SeqAIJ(): Matrix size: 8 X 8; storage space: 4 unneeded,4 used [7] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 [7] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 [1] PetscCommDuplicate(): Using internal PETSc communicator 1140850689 -2080374784 [1] PetscCommDuplicate(): returning tag 2147483643 [7] PetscCommDuplicate(): Using internal PETSc communicator 1140850689 -2080374784 [7] PetscCommDuplicate(): returning tag 2147483643 [3] PetscCommDuplicate(): Using internal PETSc communicator 1140850689 -2080374784 [3] PetscCommDuplicate(): returning tag 2147483643 [7] PetscCommDuplicate(): returning tag 2147483643 [1] PetscCommDuplicate(): returning tag 2147483643 [3] PetscCommDuplicate(): returning tag 2147483643 [0] PetscCommDuplicate(): Using internal PETSc communicator 1140850689 -2080374784 [0] PetscCommDuplicate(): returning tag 2147483643 [0] MatSetUpMultiply_MPIAIJ(): Using block index set to define scatter [0] PetscCommDuplicate(): returning tag 2147483643 [0] PetscCommDuplicate(): Using internal PETSc communicator 1140850689 -2080374784 [0] PetscCommDuplicate(): returning tag 2147483642 [0] PetscCommDuplicate(): returning tag 2147483642 [0] PetscCommDuplicate(): returning tag 2147483637 [0] VecScatterCreateCommon_PtoS(): Using blocksize 1 scatter [0] VecScatterCreate(): General case: MPI to Seq [0] MatSetOption_Inode(): Not using Inode routines due to MatSetOption(MAT_DO_NOT_USE_INODES [0] MatAssemblyEnd_SeqAIJ(): Matrix size: 11 X 8; storage space: 0 unneeded,14 used [0] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 [0] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 2 What is this "stash"? Does this mean that it is sending matrix entries between processes (b/c I should be only setting entries local to each process), or some other kind of meta-information? > -----Original Message----- > From: owner-petsc-users at mcs.anl.gov > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Barry Smith > Sent: Friday, October 26, 2007 7:54 PM > To: petsc-users at mcs.anl.gov > Subject: RE: Slow assembly > > > > The sorting does not matter. > > Under normal conditions the MatAssembly should take a > fraction of a second. The only cause that we know that slows > it down to the extreme you have is that it is sending a huge > amount of data across processes (the -info option Satish > suggested will tell us if that is true). > > Are you only call MatAssemblyBegin/End() once? You should, > don't call it > multiple times. > > The sorting is not important (in fact it takes advantage of > it automatically and does not need to be set). > > Barry > > > On Fri, 26 Oct 2007, John R. Wicks wrote: > > > I have confirmed that I am calling MatSetValues() for local > rows only > > and am only setting each value exactly once. > > > > Because of how the matrix was partitioned for another non-Petsc > > program, each partition is partitioned (by columns) into 32 blocks > > (corresponding to the row partitions). I enter the data for each > > block one row at a time, i.e., for any one SetValues call, > the entries > > are sorted by increasing column index. Does that mean I can use > > MatrixSetOption(A,MAT_COLUMNS_SORTED). Should that help? > > > > P.S.: I tried it, and it still seems to be taking quite a long time. > > > > > -----Original Message----- > > > From: owner-petsc-users at mcs.anl.gov > > > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Satish Balay > > > Sent: Friday, October 26, 2007 3:04 PM > > > To: petsc-users at mcs.anl.gov > > > Subject: Re: Slow assembly > > > > > > > > > On Fri, 26 Oct 2007, John R. Wicks wrote: > > > > > > > I am working on computing PageRank for a web scale graph > > > which uses a > > > > square matrix which is 1.2x10^8 dimensional with about 10^9 > > > entries. > > > > I have partitioned the matrix for 32 processors myself > into my own > > > > ascii format, and I know the memory allocation, so I: > > > > > > > > 1) create the matrix with "A = MatCreateMPIAIJ(*n, *n, *N, > > > *N, 0, nnz, > > > > 0, onnz)", > > > > 2) load the entries by repeatedly calling > > > > "MatSetValues(A,1,&row,links,cols,vals,INSERT_VALUES)", and > > > > > > > > 3) call MatAssemblyBegin/End. > > > > > > > > Steps 1 and 2 complete in a couple minutes, but step 3 is taking > > > > several hours. What is going on? Is there a way to speed > > > up matrix > > > > assembly? > > > > > > Are you makeing sure that you call MatGetOwnershipRange() - > > > and calling MatSetValues() for mostly local rows only? > > > > > > Also can you confirm that multiple processes [for eg: proc-0 > > > and proc-1 etc..] are not setting the same value [i.e both > > > of them calling MatSetValues(row=0,col=0)] > > > > > > Satish > > > > > > > > From knepley at gmail.com Mon Oct 29 05:53:10 2007 From: knepley at gmail.com (Matthew Knepley) Date: Mon, 29 Oct 2007 05:53:10 -0500 Subject: Slow assembly In-Reply-To: <000601c81a15$abbc0300$6b01a8c0@jwickslptp> References: <000601c81a15$abbc0300$6b01a8c0@jwickslptp> Message-ID: Yes, the stash holds entries for other processes which are sent during Assembly. Matt On 10/29/07, John R. Wicks wrote: > Although I have the malloc problem fixed, I'm wondering about the number of > messages during assembly: > [0] MatStashScatterBegin_Private(): No of messages: 0 > [0] MatAssemblyBegin_MPIAIJ(): Stash has 0 entries, uses 0 mallocs. > [1] MatStashScatterBegin_Private(): No of messages: 1 > [1] MatStashScatterBegin_Private(): Mesg_to: 0: size: 272 > [1] MatAssemblyBegin_MPIAIJ(): Stash has 33 entries, uses 0 mallocs. > [3] MatStashScatterBegin_Private(): No of messages: 3 > [3] MatStashScatterBegin_Private(): Mesg_to: 0: size: 360 > [3] MatStashScatterBegin_Private(): Mesg_to: 1: size: 328 > [3] MatStashScatterBegin_Private(): Mesg_to: 2: size: 360 > [3] MatAssemblyBegin_MPIAIJ(): Stash has 128 entries, uses 0 mallocs. > [0] MatAssemblyEnd_SeqAIJ(): Matrix size: 11 X 11; storage space: 0 > unneeded,11 used > [0] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 > [0] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 > [6] MatStashScatterBegin_Private(): No of messages: 6 > [6] MatStashScatterBegin_Private(): Mesg_to: 0: size: 184 > [6] MatStashScatterBegin_Private(): Mesg_to: 1: size: 168 > [6] MatStashScatterBegin_Private(): Mesg_to: 2: size: 184 > [6] MatStashScatterBegin_Private(): Mesg_to: 3: size: 184 > [6] MatStashScatterBegin_Private(): Mesg_to: 4: size: 152 > [6] MatStashScatterBegin_Private(): Mesg_to: 5: size: 168 > [6] MatAssemblyBegin_MPIAIJ(): Stash has 124 entries, uses 0 mallocs. > [2] MatStashScatterBegin_Private(): No of messages: 2 > [2] MatStashScatterBegin_Private(): Mesg_to: 0: size: 96 > [2] MatStashScatterBegin_Private(): Mesg_to: 1: size: 88 > [2] MatAssemblyBegin_MPIAIJ(): Stash has 21 entries, uses 0 mallocs. > [2] MatAssemblyEnd_SeqAIJ(): Matrix size: 11 X 11; storage space: 0 > unneeded,11 used > [2] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 > [2] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 > [5] MatStashScatterBegin_Private(): No of messages: 5 > [5] MatStashScatterBegin_Private(): Mesg_to: 0: size: 360 > [5] MatStashScatterBegin_Private(): Mesg_to: 1: size: 328 > [5] MatStashScatterBegin_Private(): Mesg_to: 2: size: 360 > [5] MatStashScatterBegin_Private(): Mesg_to: 3: size: 360 > [5] MatStashScatterBegin_Private(): Mesg_to: 4: size: 296 > [5] MatAssemblyBegin_MPIAIJ(): Stash has 208 entries, uses 0 mallocs. > [5] MatAssemblyEnd_SeqAIJ(): Matrix size: 10 X 10; storage space: 0 > unneeded,10 used > [5] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 > [5] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 > [6] MatAssemblyEnd_SeqAIJ(): Matrix size: 10 X 10; storage space: 0 > unneeded,10 used > [6] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 > [6] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 > [1] MatAssemblyEnd_SeqAIJ(): Matrix size: 10 X 10; storage space: 0 > unneeded,10 used > [1] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 > [1] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 > [4] MatStashScatterBegin_Private(): No of messages: 4 > [4] MatStashScatterBegin_Private(): Mesg_to: 0: size: 184 > [4] MatStashScatterBegin_Private(): Mesg_to: 1: size: 168 > [4] MatStashScatterBegin_Private(): Mesg_to: 2: size: 184 > [4] MatStashScatterBegin_Private(): Mesg_to: 3: size: 184 > [4] MatAssemblyBegin_MPIAIJ(): Stash has 86 entries, uses 0 mallocs. > [3] MatAssemblyEnd_SeqAIJ(): Matrix size: 11 X 11; storage space: 0 > unneeded,11 used > [3] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 > [3] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 > [4] MatAssemblyEnd_SeqAIJ(): Matrix size: 9 X 9; storage space: 0 unneeded,9 > used > [4] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 > [4] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 > [7] MatStashScatterBegin_Private(): No of messages: 7 > [7] MatStashScatterBegin_Private(): Mesg_to: 0: size: 184 > [7] MatStashScatterBegin_Private(): Mesg_to: 1: size: 168 > [7] MatStashScatterBegin_Private(): Mesg_to: 2: size: 184 > [7] MatStashScatterBegin_Private(): Mesg_to: 3: size: 184 > [7] MatStashScatterBegin_Private(): Mesg_to: 4: size: 152 > [7] MatStashScatterBegin_Private(): Mesg_to: 5: size: 168 > [7] MatStashScatter[6] PetscCommDuplicate(): Using internal PETSc > communicator 1140850689 -2080374784 > [6] PetscCommDuplicate(): returning tag 2147483643 > [5] PetscCommDuplicate(): Using internal PETSc communicator 1140850689 > -2080374784 > [5] PetscCommDuplicate(): returning tag 2147483643 > [2] PetscCommDuplicate(): Using internal PETSc communicator 1140850689 > -2080374784 > [2] PetscCommDuplicate(): returning tag 2147483643 > [6] PetscCommDuplicate(): returning tag 2147483643 > [5] PetscCommDuplicate(): returning tag 2147483643 > [2] PetscCommDuplicate(): returning tag 2147483643 > Begin_Private(): Mesg_to: 6: size: 168 > [7] MatAssemblyBegin_MPIAIJ(): Stash has 144 entries, uses 0 mallocs. > [7] MatAssemblyEnd_SeqAIJ(): Matrix size: 8 X 8; storage space: 4 unneeded,4 > used > [7] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 > [7] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 > [1] PetscCommDuplicate(): Using internal PETSc communicator 1140850689 > -2080374784 > [1] PetscCommDuplicate(): returning tag 2147483643 > [7] PetscCommDuplicate(): Using internal PETSc communicator 1140850689 > -2080374784 > [7] PetscCommDuplicate(): returning tag 2147483643 > [3] PetscCommDuplicate(): Using internal PETSc communicator 1140850689 > -2080374784 > [3] PetscCommDuplicate(): returning tag 2147483643 > [7] PetscCommDuplicate(): returning tag 2147483643 > [1] PetscCommDuplicate(): returning tag 2147483643 > [3] PetscCommDuplicate(): returning tag 2147483643 > [0] PetscCommDuplicate(): Using internal PETSc communicator 1140850689 > -2080374784 > [0] PetscCommDuplicate(): returning tag 2147483643 > [0] MatSetUpMultiply_MPIAIJ(): Using block index set to define scatter > [0] PetscCommDuplicate(): returning tag 2147483643 > [0] PetscCommDuplicate(): Using internal PETSc communicator 1140850689 > -2080374784 > [0] PetscCommDuplicate(): returning tag 2147483642 > [0] PetscCommDuplicate(): returning tag 2147483642 > [0] PetscCommDuplicate(): returning tag 2147483637 > [0] VecScatterCreateCommon_PtoS(): Using blocksize 1 scatter > [0] VecScatterCreate(): General case: MPI to Seq > [0] MatSetOption_Inode(): Not using Inode routines due to > MatSetOption(MAT_DO_NOT_USE_INODES > [0] MatAssemblyEnd_SeqAIJ(): Matrix size: 11 X 8; storage space: 0 > unneeded,14 used > [0] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 > [0] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 2 > > What is this "stash"? Does this mean that it is sending matrix entries > between processes (b/c I should be only setting entries local to each > process), or some other kind of meta-information? > > > -----Original Message----- > > From: owner-petsc-users at mcs.anl.gov > > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Barry Smith > > Sent: Friday, October 26, 2007 7:54 PM > > To: petsc-users at mcs.anl.gov > > Subject: RE: Slow assembly > > > > > > > > The sorting does not matter. > > > > Under normal conditions the MatAssembly should take a > > fraction of a second. The only cause that we know that slows > > it down to the extreme you have is that it is sending a huge > > amount of data across processes (the -info option Satish > > suggested will tell us if that is true). > > > > Are you only call MatAssemblyBegin/End() once? You should, > > don't call it > > multiple times. > > > > The sorting is not important (in fact it takes advantage of > > it automatically and does not need to be set). > > > > Barry > > > > > > On Fri, 26 Oct 2007, John R. Wicks wrote: > > > > > I have confirmed that I am calling MatSetValues() for local > > rows only > > > and am only setting each value exactly once. > > > > > > Because of how the matrix was partitioned for another non-Petsc > > > program, each partition is partitioned (by columns) into 32 blocks > > > (corresponding to the row partitions). I enter the data for each > > > block one row at a time, i.e., for any one SetValues call, > > the entries > > > are sorted by increasing column index. Does that mean I can use > > > MatrixSetOption(A,MAT_COLUMNS_SORTED). Should that help? > > > > > > P.S.: I tried it, and it still seems to be taking quite a long time. > > > > > > > -----Original Message----- > > > > From: owner-petsc-users at mcs.anl.gov > > > > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Satish Balay > > > > Sent: Friday, October 26, 2007 3:04 PM > > > > To: petsc-users at mcs.anl.gov > > > > Subject: Re: Slow assembly > > > > > > > > > > > > On Fri, 26 Oct 2007, John R. Wicks wrote: > > > > > > > > > I am working on computing PageRank for a web scale graph > > > > which uses a > > > > > square matrix which is 1.2x10^8 dimensional with about 10^9 > > > > entries. > > > > > I have partitioned the matrix for 32 processors myself > > into my own > > > > > ascii format, and I know the memory allocation, so I: > > > > > > > > > > 1) create the matrix with "A = MatCreateMPIAIJ(*n, *n, *N, > > > > *N, 0, nnz, > > > > > 0, onnz)", > > > > > 2) load the entries by repeatedly calling > > > > > "MatSetValues(A,1,&row,links,cols,vals,INSERT_VALUES)", and > > > > > > > > > > 3) call MatAssemblyBegin/End. > > > > > > > > > > Steps 1 and 2 complete in a couple minutes, but step 3 is taking > > > > > several hours. What is going on? Is there a way to speed > > > > up matrix > > > > > assembly? > > > > > > > > Are you makeing sure that you call MatGetOwnershipRange() - > > > > and calling MatSetValues() for mostly local rows only? > > > > > > > > Also can you confirm that multiple processes [for eg: proc-0 > > > > and proc-1 etc..] are not setting the same value [i.e both > > > > of them calling MatSetValues(row=0,col=0)] > > > > > > > > Satish > > > > > > > > > > > > > > -- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. -- Norbert Wiener From jwicks at cs.brown.edu Mon Oct 29 06:15:44 2007 From: jwicks at cs.brown.edu (John R. Wicks) Date: Mon, 29 Oct 2007 07:15:44 -0400 Subject: Precision of output In-Reply-To: Message-ID: <000201c81a1d$0dac0450$6b01a8c0@jwickslptp> How can I set the precision of my output? I use an ASCII viewer to write the computed vector to a file and many of the entries should on the order of 10^-9, but they seem to be printing out at 0. How can I get it to print to higher precision? From bsmith at mcs.anl.gov Mon Oct 29 06:55:09 2007 From: bsmith at mcs.anl.gov (Barry Smith) Date: Mon, 29 Oct 2007 06:55:09 -0500 (CDT) Subject: Precision of output In-Reply-To: <000201c81a1d$0dac0450$6b01a8c0@jwickslptp> References: <000201c81a1d$0dac0450$6b01a8c0@jwickslptp> Message-ID: -mat_view_matlab or just before calling the MatView() call ierr = PetscViewerPushFormat(viewer,PETSC_VIEWER_ASCII_MATLAB);CHKERRQ(ierr); On Mon, 29 Oct 2007, John R. Wicks wrote: > How can I set the precision of my output? I use an ASCII viewer to write > the computed vector to a file and many of the entries should on the order of > 10^-9, but they seem to be printing out at 0. How can I get it to print to > higher precision? > > From bsmith at mcs.anl.gov Mon Oct 29 06:56:42 2007 From: bsmith at mcs.anl.gov (Barry Smith) Date: Mon, 29 Oct 2007 06:56:42 -0500 (CDT) Subject: Slow assembly In-Reply-To: <000601c81a15$abbc0300$6b01a8c0@jwickslptp> References: <000601c81a15$abbc0300$6b01a8c0@jwickslptp> Message-ID: You are assigning some matrix values on a process that is different fromm the one where it will "live". In general you want to generate 90%+ of matrix elements on the process where they live, otherwise the MatAssembly will be slow. Barry On Mon, 29 Oct 2007, John R. Wicks wrote: > Although I have the malloc problem fixed, I'm wondering about the number of > messages during assembly: > [0] MatStashScatterBegin_Private(): No of messages: 0 > [0] MatAssemblyBegin_MPIAIJ(): Stash has 0 entries, uses 0 mallocs. > [1] MatStashScatterBegin_Private(): No of messages: 1 > [1] MatStashScatterBegin_Private(): Mesg_to: 0: size: 272 > [1] MatAssemblyBegin_MPIAIJ(): Stash has 33 entries, uses 0 mallocs. > [3] MatStashScatterBegin_Private(): No of messages: 3 > [3] MatStashScatterBegin_Private(): Mesg_to: 0: size: 360 > [3] MatStashScatterBegin_Private(): Mesg_to: 1: size: 328 > [3] MatStashScatterBegin_Private(): Mesg_to: 2: size: 360 > [3] MatAssemblyBegin_MPIAIJ(): Stash has 128 entries, uses 0 mallocs. > [0] MatAssemblyEnd_SeqAIJ(): Matrix size: 11 X 11; storage space: 0 > unneeded,11 used > [0] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 > [0] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 > [6] MatStashScatterBegin_Private(): No of messages: 6 > [6] MatStashScatterBegin_Private(): Mesg_to: 0: size: 184 > [6] MatStashScatterBegin_Private(): Mesg_to: 1: size: 168 > [6] MatStashScatterBegin_Private(): Mesg_to: 2: size: 184 > [6] MatStashScatterBegin_Private(): Mesg_to: 3: size: 184 > [6] MatStashScatterBegin_Private(): Mesg_to: 4: size: 152 > [6] MatStashScatterBegin_Private(): Mesg_to: 5: size: 168 > [6] MatAssemblyBegin_MPIAIJ(): Stash has 124 entries, uses 0 mallocs. > [2] MatStashScatterBegin_Private(): No of messages: 2 > [2] MatStashScatterBegin_Private(): Mesg_to: 0: size: 96 > [2] MatStashScatterBegin_Private(): Mesg_to: 1: size: 88 > [2] MatAssemblyBegin_MPIAIJ(): Stash has 21 entries, uses 0 mallocs. > [2] MatAssemblyEnd_SeqAIJ(): Matrix size: 11 X 11; storage space: 0 > unneeded,11 used > [2] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 > [2] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 > [5] MatStashScatterBegin_Private(): No of messages: 5 > [5] MatStashScatterBegin_Private(): Mesg_to: 0: size: 360 > [5] MatStashScatterBegin_Private(): Mesg_to: 1: size: 328 > [5] MatStashScatterBegin_Private(): Mesg_to: 2: size: 360 > [5] MatStashScatterBegin_Private(): Mesg_to: 3: size: 360 > [5] MatStashScatterBegin_Private(): Mesg_to: 4: size: 296 > [5] MatAssemblyBegin_MPIAIJ(): Stash has 208 entries, uses 0 mallocs. > [5] MatAssemblyEnd_SeqAIJ(): Matrix size: 10 X 10; storage space: 0 > unneeded,10 used > [5] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 > [5] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 > [6] MatAssemblyEnd_SeqAIJ(): Matrix size: 10 X 10; storage space: 0 > unneeded,10 used > [6] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 > [6] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 > [1] MatAssemblyEnd_SeqAIJ(): Matrix size: 10 X 10; storage space: 0 > unneeded,10 used > [1] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 > [1] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 > [4] MatStashScatterBegin_Private(): No of messages: 4 > [4] MatStashScatterBegin_Private(): Mesg_to: 0: size: 184 > [4] MatStashScatterBegin_Private(): Mesg_to: 1: size: 168 > [4] MatStashScatterBegin_Private(): Mesg_to: 2: size: 184 > [4] MatStashScatterBegin_Private(): Mesg_to: 3: size: 184 > [4] MatAssemblyBegin_MPIAIJ(): Stash has 86 entries, uses 0 mallocs. > [3] MatAssemblyEnd_SeqAIJ(): Matrix size: 11 X 11; storage space: 0 > unneeded,11 used > [3] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 > [3] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 > [4] MatAssemblyEnd_SeqAIJ(): Matrix size: 9 X 9; storage space: 0 unneeded,9 > used > [4] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 > [4] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 > [7] MatStashScatterBegin_Private(): No of messages: 7 > [7] MatStashScatterBegin_Private(): Mesg_to: 0: size: 184 > [7] MatStashScatterBegin_Private(): Mesg_to: 1: size: 168 > [7] MatStashScatterBegin_Private(): Mesg_to: 2: size: 184 > [7] MatStashScatterBegin_Private(): Mesg_to: 3: size: 184 > [7] MatStashScatterBegin_Private(): Mesg_to: 4: size: 152 > [7] MatStashScatterBegin_Private(): Mesg_to: 5: size: 168 > [7] MatStashScatter[6] PetscCommDuplicate(): Using internal PETSc > communicator 1140850689 -2080374784 > [6] PetscCommDuplicate(): returning tag 2147483643 > [5] PetscCommDuplicate(): Using internal PETSc communicator 1140850689 > -2080374784 > [5] PetscCommDuplicate(): returning tag 2147483643 > [2] PetscCommDuplicate(): Using internal PETSc communicator 1140850689 > -2080374784 > [2] PetscCommDuplicate(): returning tag 2147483643 > [6] PetscCommDuplicate(): returning tag 2147483643 > [5] PetscCommDuplicate(): returning tag 2147483643 > [2] PetscCommDuplicate(): returning tag 2147483643 > Begin_Private(): Mesg_to: 6: size: 168 > [7] MatAssemblyBegin_MPIAIJ(): Stash has 144 entries, uses 0 mallocs. > [7] MatAssemblyEnd_SeqAIJ(): Matrix size: 8 X 8; storage space: 4 unneeded,4 > used > [7] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 > [7] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 1 > [1] PetscCommDuplicate(): Using internal PETSc communicator 1140850689 > -2080374784 > [1] PetscCommDuplicate(): returning tag 2147483643 > [7] PetscCommDuplicate(): Using internal PETSc communicator 1140850689 > -2080374784 > [7] PetscCommDuplicate(): returning tag 2147483643 > [3] PetscCommDuplicate(): Using internal PETSc communicator 1140850689 > -2080374784 > [3] PetscCommDuplicate(): returning tag 2147483643 > [7] PetscCommDuplicate(): returning tag 2147483643 > [1] PetscCommDuplicate(): returning tag 2147483643 > [3] PetscCommDuplicate(): returning tag 2147483643 > [0] PetscCommDuplicate(): Using internal PETSc communicator 1140850689 > -2080374784 > [0] PetscCommDuplicate(): returning tag 2147483643 > [0] MatSetUpMultiply_MPIAIJ(): Using block index set to define scatter > [0] PetscCommDuplicate(): returning tag 2147483643 > [0] PetscCommDuplicate(): Using internal PETSc communicator 1140850689 > -2080374784 > [0] PetscCommDuplicate(): returning tag 2147483642 > [0] PetscCommDuplicate(): returning tag 2147483642 > [0] PetscCommDuplicate(): returning tag 2147483637 > [0] VecScatterCreateCommon_PtoS(): Using blocksize 1 scatter > [0] VecScatterCreate(): General case: MPI to Seq > [0] MatSetOption_Inode(): Not using Inode routines due to > MatSetOption(MAT_DO_NOT_USE_INODES > [0] MatAssemblyEnd_SeqAIJ(): Matrix size: 11 X 8; storage space: 0 > unneeded,14 used > [0] MatAssemblyEnd_SeqAIJ(): Number of mallocs during MatSetValues() is 0 > [0] MatAssemblyEnd_SeqAIJ(): Maximum nonzeros in any row is 2 > > What is this "stash"? Does this mean that it is sending matrix entries > between processes (b/c I should be only setting entries local to each > process), or some other kind of meta-information? > > > -----Original Message----- > > From: owner-petsc-users at mcs.anl.gov > > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Barry Smith > > Sent: Friday, October 26, 2007 7:54 PM > > To: petsc-users at mcs.anl.gov > > Subject: RE: Slow assembly > > > > > > > > The sorting does not matter. > > > > Under normal conditions the MatAssembly should take a > > fraction of a second. The only cause that we know that slows > > it down to the extreme you have is that it is sending a huge > > amount of data across processes (the -info option Satish > > suggested will tell us if that is true). > > > > Are you only call MatAssemblyBegin/End() once? You should, > > don't call it > > multiple times. > > > > The sorting is not important (in fact it takes advantage of > > it automatically and does not need to be set). > > > > Barry > > > > > > On Fri, 26 Oct 2007, John R. Wicks wrote: > > > > > I have confirmed that I am calling MatSetValues() for local > > rows only > > > and am only setting each value exactly once. > > > > > > Because of how the matrix was partitioned for another non-Petsc > > > program, each partition is partitioned (by columns) into 32 blocks > > > (corresponding to the row partitions). I enter the data for each > > > block one row at a time, i.e., for any one SetValues call, > > the entries > > > are sorted by increasing column index. Does that mean I can use > > > MatrixSetOption(A,MAT_COLUMNS_SORTED). Should that help? > > > > > > P.S.: I tried it, and it still seems to be taking quite a long time. > > > > > > > -----Original Message----- > > > > From: owner-petsc-users at mcs.anl.gov > > > > [mailto:owner-petsc-users at mcs.anl.gov] On Behalf Of Satish Balay > > > > Sent: Friday, October 26, 2007 3:04 PM > > > > To: petsc-users at mcs.anl.gov > > > > Subject: Re: Slow assembly > > > > > > > > > > > > On Fri, 26 Oct 2007, John R. Wicks wrote: > > > > > > > > > I am working on computing PageRank for a web scale graph > > > > which uses a > > > > > square matrix which is 1.2x10^8 dimensional with about 10^9 > > > > entries. > > > > > I have partitioned the matrix for 32 processors myself > > into my own > > > > > ascii format, and I know the memory allocation, so I: > > > > > > > > > > 1) create the matrix with "A = MatCreateMPIAIJ(*n, *n, *N, > > > > *N, 0, nnz, > > > > > 0, onnz)", > > > > > 2) load the entries by repeatedly calling > > > > > "MatSetValues(A,1,&row,links,cols,vals,INSERT_VALUES)", and > > > > > > > > > > 3) call MatAssemblyBegin/End. > > > > > > > > > > Steps 1 and 2 complete in a couple minutes, but step 3 is taking > > > > > several hours. What is going on? Is there a way to speed > > > > up matrix > > > > > assembly? > > > > > > > > Are you makeing sure that you call MatGetOwnershipRange() - > > > > and calling MatSetValues() for mostly local rows only? > > > > > > > > Also can you confirm that multiple processes [for eg: proc-0 > > > > and proc-1 etc..] are not setting the same value [i.e both > > > > of them calling MatSetValues(row=0,col=0)] > > > > > > > > Satish > > > > > > > > > > > > > > From keita at cray.com Mon Oct 29 15:37:57 2007 From: keita at cray.com (Keita Teranishi) Date: Mon, 29 Oct 2007 15:37:57 -0500 Subject: Error in LCD Message-ID: <925346A443D4E340BEB20248BAFCDBDF02D8501F@CFEVS1-IP.americas.cray.com> Hi, I am having a trouble with LCD in KSP module. I got following error messages as I executed src/ksp/ksp/example/tutorials/ex10. There is no error message when running much simpler examples such as ex2. Regards, keita $ ex10 -f0 ./xerox2000c3.spd -f1 ./xerox2000c3.spd -ksp_type lcd [0]PETSC ERROR: --------------------- Error Message ------------------------------------ [0]PETSC ERROR: Null argument, when expecting valid pointer! [0]PETSC ERROR: Null Object: Parameter # 1! [0]PETSC ERROR: ------------------------------------------------------------------------ : : : : [0]PETSC ERROR: ------------------------------------------------------------------------ [0]PETSC ERROR: VecDuplicateVecs() line 533 in src/vec/vec/interface/vector.c [0]PETSC ERROR: KSPSetUp_LCD() line 48 in src/ksp/ksp/impls/lcd/lcd.c [0]PETSC ERROR: KSPSetUp() line 198 in src/ksp/ksp/interface/itfunc.c [0]PETSC ERROR: main() line 293 in src/ksp/ksp/examples/tutorials/ex10.c [unset]: aborting job: application called MPI_Abort(MPI_COMM_WORLD, 85) - process 0 ================================ Keita Teranishi Math Software Group Cray Inc. keita at cray.com ================================ -------------- next part -------------- An HTML attachment was scrubbed... URL: From dalcinl at gmail.com Mon Oct 29 15:49:09 2007 From: dalcinl at gmail.com (Lisandro Dalcin) Date: Mon, 29 Oct 2007 17:49:09 -0300 Subject: Error in LCD In-Reply-To: <925346A443D4E340BEB20248BAFCDBDF02D8501F@CFEVS1-IP.americas.cray.com> References: <925346A443D4E340BEB20248BAFCDBDF02D8501F@CFEVS1-IP.americas.cray.com> Message-ID: Are you calling KSPSetUp() before calling KSPSolve() ?? That should work, but I believe some KSP methods are going to have problem with this. On 10/29/07, Keita Teranishi wrote: > Hi, > I am having a trouble with LCD in KSP module. I got following error > messages as I executed src/ksp/ksp/example/tutorials/ex10. > There is no error message when running much simpler examples such as ex2. > > > > Regards, > > > > keita $ ex10 -f0 ./xerox2000c3.spd -f1 ./xerox2000c3.spd -ksp_type lcd > > [0]PETSC ERROR: --------------------- Error Message > ------------------------------------ > > [0]PETSC ERROR: Null argument, when expecting valid pointer! > > [0]PETSC ERROR: Null Object: Parameter # 1! > > [0]PETSC ERROR: > ------------------------------------------------------------------------ > > : > > : > > : > > : > > [0]PETSC ERROR: > ------------------------------------------------------------------------ > > [0]PETSC ERROR: VecDuplicateVecs() line 533 in > src/vec/vec/interface/vector.c > > [0]PETSC ERROR: KSPSetUp_LCD() line 48 in src/ksp/ksp/impls/lcd/lcd.c > > [0]PETSC ERROR: KSPSetUp() line 198 in src/ksp/ksp/interface/itfunc.c > > [0]PETSC ERROR: main() line 293 in > src/ksp/ksp/examples/tutorials/ex10.c > > [unset]: aborting job: > > application called MPI_Abort(MPI_COMM_WORLD, 85) - process 0 > > > > ================================ > Keita Teranishi > Math Software Group > Cray Inc. > keita at cray.com > ================================ > > -- Lisandro Dalc?n --------------- Centro Internacional de M?todos Computacionales en Ingenier?a (CIMEC) Instituto de Desarrollo Tecnol?gico para la Industria Qu?mica (INTEC) Consejo Nacional de Investigaciones Cient?ficas y T?cnicas (CONICET) PTLC - G?emes 3450, (3000) Santa Fe, Argentina Tel/Fax: +54-(0)342-451.1594 From bsmith at mcs.anl.gov Mon Oct 29 15:51:32 2007 From: bsmith at mcs.anl.gov (Barry Smith) Date: Mon, 29 Oct 2007 15:51:32 -0500 (CDT) Subject: Error in LCD In-Reply-To: <925346A443D4E340BEB20248BAFCDBDF02D8501F@CFEVS1-IP.americas.cray.com> References: <925346A443D4E340BEB20248BAFCDBDF02D8501F@CFEVS1-IP.americas.cray.com> Message-ID: Keita, Thanks for reporting the problem. It is our error, we changed the management of work vectors so KSPSetUp() could be used before KSPSolve() but it was not updated correctly in this file. I have pushed a fix to the release repository and it will be in our next patch. For now you simply change the two lines from ierr = VecDuplicateVecs(ksp->vec_rhs,restart+1,&lcd->P); CHKERRQ(ierr); ierr = VecDuplicateVecs(ksp->vec_rhs, restart + 1, &lcd->Q); CHKERRQ(ierr); to ierr = VecDuplicateVecs(ksp->work[0],restart+1,&lcd->P); CHKERRQ(ierr); ierr = VecDuplicateVecs(ksp->work[0], restart + 1, &lcd->Q); CHKERRQ(ierr); Please report these bugs to petsc-maint at mcs.anl.gov instead, thanks Barry On Mon, 29 Oct 2007, Keita Teranishi wrote: > Hi, > > > > I am having a trouble with LCD in KSP module. I got following error messages as I executed src/ksp/ksp/example/tutorials/ex10. There is no error message when running much simpler examples such as ex2. > > > > Regards, > > > > keita $ ex10 -f0 ./xerox2000c3.spd -f1 ./xerox2000c3.spd -ksp_type lcd > > [0]PETSC ERROR: --------------------- Error Message ------------------------------------ > > [0]PETSC ERROR: Null argument, when expecting valid pointer! > > [0]PETSC ERROR: Null Object: Parameter # 1! > > [0]PETSC ERROR: ------------------------------------------------------------------------ > > : > > : > > : > > : > > [0]PETSC ERROR: ------------------------------------------------------------------------ > > [0]PETSC ERROR: VecDuplicateVecs() line 533 in src/vec/vec/interface/vector.c > > [0]PETSC ERROR: KSPSetUp_LCD() line 48 in src/ksp/ksp/impls/lcd/lcd.c > > [0]PETSC ERROR: KSPSetUp() line 198 in src/ksp/ksp/interface/itfunc.c > > [0]PETSC ERROR: main() line 293 in src/ksp/ksp/examples/tutorials/ex10.c > > [unset]: aborting job: > > application called MPI_Abort(MPI_COMM_WORLD, 85) - process 0 > > > > ================================ > Keita Teranishi > Math Software Group > Cray Inc. > keita at cray.com > ================================ > > > > From z.sheng at ewi.tudelft.nl Wed Oct 31 11:23:44 2007 From: z.sheng at ewi.tudelft.nl (Zhifeng Sheng) Date: Wed, 31 Oct 2007 17:23:44 +0100 Subject: PCFactorSetLevels does not work?! In-Reply-To: References: <925346A443D4E340BEB20248BAFCDBDF02D8501F@CFEVS1-IP.americas.cray.com> Message-ID: <4728AC10.4040500@ewi.tudelft.nl> Dear all I have a weird problem. In my programme, I try to use ICC(K), while I set the factor level with PCFactorSetLevels(prec, 4); However this does not work. Then I run the executable with the option *-pc_factor_levels 4, it works... Can someone tell me why this is happenning? Thank you Best regards Zhifeng Sheng * From dalcinl at gmail.com Wed Oct 31 12:25:32 2007 From: dalcinl at gmail.com (Lisandro Dalcin) Date: Wed, 31 Oct 2007 14:25:32 -0300 Subject: PCFactorSetLevels does not work?! In-Reply-To: <4728AC10.4040500@ewi.tudelft.nl> References: <925346A443D4E340BEB20248BAFCDBDF02D8501F@CFEVS1-IP.americas.cray.com> <4728AC10.4040500@ewi.tudelft.nl> Message-ID: Did you call PCSetType(prec,PCICC) or PCSetFromOptions(prec) before calling PCFactorSetLevels(prec, 4) ?? If not, that call would be ignored. On 10/31/07, Zhifeng Sheng wrote: > Dear all > > I have a weird problem. > > In my programme, I try to use ICC(K), while I set the factor level with > > PCFactorSetLevels(prec, 4); > > However this does not work. > > Then I run the executable with the option *-pc_factor_levels 4, it works... > > Can someone tell me why this is happenning? > > Thank you > > Best regards > Zhifeng Sheng > * > > -- Lisandro Dalc?n --------------- Centro Internacional de M?todos Computacionales en Ingenier?a (CIMEC) Instituto de Desarrollo Tecnol?gico para la Industria Qu?mica (INTEC) Consejo Nacional de Investigaciones Cient?ficas y T?cnicas (CONICET) PTLC - G?emes 3450, (3000) Santa Fe, Argentina Tel/Fax: +54-(0)342-451.1594 From bsmith at mcs.anl.gov Wed Oct 31 12:28:26 2007 From: bsmith at mcs.anl.gov (Barry Smith) Date: Wed, 31 Oct 2007 12:28:26 -0500 (CDT) Subject: PCFactorSetLevels does not work?! In-Reply-To: <4728AC10.4040500@ewi.tudelft.nl> References: <925346A443D4E340BEB20248BAFCDBDF02D8501F@CFEVS1-IP.americas.cray.com> <4728AC10.4040500@ewi.tudelft.nl> Message-ID: Are you calling the PCFactorSetLevels() on the prec where the ICC is being used. For example if your * below means -sub_pc_factor_levels then you need to use PCBJacobiGetSubKSP() then get the PC out of that that subksp and call the PCFactorSetLevels() on that Barry On Wed, 31 Oct 2007, Zhifeng Sheng wrote: > Dear all > > I have a weird problem. > > In my programme, I try to use ICC(K), while I set the factor level with > > PCFactorSetLevels(prec, 4); > > However this does not work. > > Then I run the executable with the option *-pc_factor_levels 4, it works... > Can someone tell me why this is happenning? > > Thank you > > Best regards > Zhifeng Sheng > * > > From jinzishuai at yahoo.com Wed Oct 31 18:45:00 2007 From: jinzishuai at yahoo.com (Shi Jin) Date: Wed, 31 Oct 2007 16:45:00 -0700 (PDT) Subject: memory usage of the unstructured grid Message-ID: <830561.29161.qm@web36201.mail.mud.yahoo.com> Hi, I have observed that the memory usage of the petsc mesh is much higher than my previous code, if both were to be run serially. For example, for a simple cubic box with 750,000 tetrahedral elements, my old code takes about 200MB for the whole array, including all the mappings required for later use such as the inverse connectivity table. For the same mesh, my PETSc code takes about 4GB for the mesh alone. The same can be found in the provided examples. I made a few changes to the navierStokes code to output the virtual memory usage and got ./navierStokes -dim 3 -generate -structured 0 -refinement_limit 1e-6 109,283 elements, 139,030 edges , 21,523 vertexes [0]:after mesh created:mem=574.46 MB This is consistent with my Petsc code. I understand that for the mesh to scale in parallel, extra information needs to be stored. But the current cost seems too expensive. I am wondering if there is a way to cut the memory usage for the mesh. Thank you very much. Shi __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com -------------- next part -------------- An HTML attachment was scrubbed... URL: From jens.madsen at risoe.dk Wed Oct 31 09:38:28 2007 From: jens.madsen at risoe.dk (jens.madsen at risoe.dk) Date: Wed, 31 Oct 2007 15:38:28 +0100 Subject: Larger stencils in DA Message-ID: Hi Do you have examples where stencils not fitting the DA types (box and star) are being used? Are there any plans of introducing a stencil width into DA? Kind Regards Jens MAdsen -------------- next part -------------- An HTML attachment was scrubbed... URL: