[petsc-users] PetscSplitOwnerShip()
Matthew Knepley
knepley at gmail.com
Wed Mar 10 10:28:13 CST 2010
Why are you loading the vector twice? Shouldn't you just use
VecLoadIntoVector()?
Matt
On Wed, Mar 10, 2010 at 9:50 AM, (Rebecca) Xuefei YUAN
<xy2102 at columbia.edu>wrote:
> Dear Barry and Matt,
>
> Thanks very much for the reply.
>
> I am a little bit confused about binary write and read as an input and
> output method.
>
> For example, I have code1 and code2, where code1 will take code2's output
> as an input.
>
> In code2, the following routine is used to save the solution as a binary
> file:
>
>
> #undef __FUNCT__
> #define __FUNCT__ "DumpSolutionToMatlab"
> PetscErrorCode DumpSolutionToMatlab (DMMG*dmmg, char * fn)
> {
> DALocalInfo info;
> PetscViewer viewer;
> PetscFunctionBegin;
>
> ierr = DAGetLocalInfo (DMMGGetDA(dmmg),&info);CHKERRQ(ierr);
> sprintf(fileName,
> "ff_atwqt2ff_tx%i_ty%i_x%i_y%i_nl%i_gt%i_ot%i_s%i.dat",info.mx
> ,info.my,parameters->mxfield,parameters->myfield,parameters->numberOfLevels,(PetscInt)(parameters->currentTime*10000),parameters->timeAccuracyOrder,(PetscInt)(parameters->smoothFactor*100));
> ierr =
> PetscViewerBinaryOpen(PETSC_COMM_WORLD,fileName,FILE_MODE_WRITE,&viewer);CHKERRQ(ierr);
> ierr =
> PetscObjectSetName((PetscObject)dmmg[DMMGGetLevels(dmmg)-1]->x,"ff");
> ierr = VecView(dmmg[DMMGGetLevels(dmmg)-1]->x,viewer);CHKERRQ(ierr);
> ierr = PetscViewerDestroy (viewer); CHKERRQ (ierr);
>
> PetscFunctionReturn(0);
> }
>
>
> In code1, the following routine is used to read the solution to a Vec named
> FIELD:
>
> #undef __FUNCT__
> #define __FUNCT__ "FormInitialGuess_physical"
> PetscErrorCode FormInitialGuess_physical(DMMG dmmg, Vec X)
> {
> Vec FIELD;
> Field **field;
> DA da,da2_4;
> PetscViewer viewer;
> DALocalInfo info,info2;
>
> PetscFunctionBegin;
>
> ierr =
> DMCompositeGetEntries((DMComposite)(dmmg->dm),&da,PETSC_IGNORE);CHKERRQ(ierr);
> ierr = DAGetLocalInfo(da,&info);CHKERRQ(ierr);
> ierr = DACreate2d(PETSC_COMM_WORLD,DA_NONPERIODIC,DA_STENCIL_BOX,
> (info.mx-1-parameters->abandonNumber),
> (info.my-1-parameters->abandonNumber), PETSC_DECIDE, PETSC_DECIDE, 4, 2,
> PETSC_NULL, PETSC_NULL, &da2_4);CHKERRQ(ierr);
> ierr = DAGetLocalInfo(da2_4,&info2);CHKERRQ(ierr);
> sprintf(fileName, "ff_twqt2ff_tx%i_ty%i_x%i_y%i_nl%i_gt%i_ot%i_s%i.dat",
> info2.mx
> ,info2.my,parameters->mxgrid-1,parameters->mygrid-1,parameters->numberOfLevels,(PetscInt)(parameters->timeToGenerateGrid*10000),parameters->timeAccuracyOrder,(PetscInt)(parameters->smoothFactor*100));
>
> PetscViewerBinaryOpen(PETSC_COMM_WORLD,fileName,FILE_MODE_READ,&viewer);
> VecLoad(viewer,PETSC_NULL,&FIELD);
> VecLoadIntoVector(viewer,FIELD);
> ierr = DAVecGetArray(da2_4,FIELD,&field);CHKERRQ(ierr);
>
> ierr = DAVecRestoreArray(da2_4,FIELD,&field);CHKERRQ(ierr);
> ierr = VecDestroy(FIELD);CHKERRQ(ierr);
> ierr = DADestroy(da2_4);CHKERRQ(ierr);
> ierr = PetscViewerDestroy(viewer);CHKERRQ(ierr);
>
> PetscFunctionReturn(0);
> }
>
> When I call
> VecLoad(viewer,PETSC_NULL,&FIELD);
> VecLoadIntoVector(viewer,FIELD);
>
> there is an error in proc0:
> [0]PETSC ERROR: PetscBinaryRead() line 251 in src/sys/fileio/sysio.c Read
> past end of file
>
> In gdb, it shows that
>
> Program received signal SIGABRT, Aborted.
> [Switching to Thread 0xb7c7c6b0 (LWP 598)]
> 0xb7f4c410 in __kernel_vsyscall ()
> (gdb) where
> #0 0xb7f4c410 in __kernel_vsyscall ()
> #1 0xb7ccc085 in raise () from /lib/tls/i686/cmov/libc.so.6
> #2 0xb7ccda01 in abort () from /lib/tls/i686/cmov/libc.so.6
> #3 0x08733fe9 in PetscAbortErrorHandler (line=251,
> fun=0x886a9c9 "PetscBinaryRead", file=0x886a942 "sysio.c",
> dir=0x886a94a "src/sys/fileio/", n=66, p=1,
> mess=0xbffdbe64 "Read past end of file", ctx=0x0) at errabort.c:62
> #4 0x086a8f5a in PetscError (line=251, func=0x886a9c9 "PetscBinaryRead",
> file=0x886a942 "sysio.c", dir=0x886a94a "src/sys/fileio/", n=66, p=1,
> mess=0x886a9e6 "Read past end of file") at err.c:482
> #5 0x086b73a2 in PetscBinaryRead (fd=9, p=0xbffdc764, n=1, type=PETSC_INT)
> at sysio.c:251
> #6 0x085d8dee in VecLoadIntoVector_Binary (viewer=0x89f2870,
> vec=0x89f4a00)
> at vecio.c:445
> #7 0x085d9e76 in VecLoadIntoVector_Default (viewer=0x89f2870,
> vec=0x89f4a00)
> at vecio.c:514
> #8 0x085e9d2c in VecLoadIntoVector (viewer=0x89f2870, vec=0x89f4a00)
> at vector.c:1031
> #9 0x0804f158 in FormInitialGuess_physical (dmmg=0x8999b30, X=0x898b3c0)
> at vecviewload_out.c:391
> #10 0x08052d05 in DMMGSolve (dmmg=0x89999e0) at damg.c:307
> #11 0x0804d479 in main (argc=Cannot access memory at address 0x256
> ) at vecviewload_out.c:186
>
> When I call
> VecLoad(viewer,PETSC_NULL,&FIELD);
> // VecLoadIntoVector(viewer,FIELD);
> [0]PETSC ERROR: [1]PETSC ERROR: DAVecGetArray() line 53 in
> src/dm/da/src/dagetarray.c Vector local size 32 is not compatible with DA
> local sizes 36 100
>
> [2]PETSC ERROR: [3]PETSC ERROR: DAVecGetArray() line 53 in
> src/dm/da/src/dagetarray.c Vector local size 28 is not compatible with DA
> local sizes 24 80
>
> DAVecGetArray() line 53 in src/dm/da/src/dagetarray.c Vector local size 32
> is not compatible with DA local sizes 36 100
>
> DAVecGetArray() line 53 in src/dm/da/src/dagetarray.c Vector local size 28
> is not compatible with DA local sizes 24 80
>
> in gdb:
>
> Program received signal SIGABRT, Aborted.
> [Switching to Thread 0xb7c4f6b0 (LWP 840)]
> 0xb7f1f410 in __kernel_vsyscall ()
> (gdb) where
> #0 0xb7f1f410 in __kernel_vsyscall ()
> #1 0xb7c9f085 in raise () from /lib/tls/i686/cmov/libc.so.6
> #2 0xb7ca0a01 in abort () from /lib/tls/i686/cmov/libc.so.6
> #3 0x08733fd1 in PetscAbortErrorHandler (line=53,
> fun=0x883be60 "DAVecGetArray", file=0x883be6e "dagetarray.c",
> dir=0x883be7b "src/dm/da/src/", n=75, p=1,
> mess=0xbffcbd54 "Vector local size 32 is not compatible with DA local
> sizes 36 100\n", ctx=0x0) at errabort.c:62
> #4 0x086a8f42 in PetscError (line=53, func=0x883be60 "DAVecGetArray",
> file=0x883be6e "dagetarray.c", dir=0x883be7b "src/dm/da/src/", n=75,
> p=1,
> mess=0x883be8c "Vector local size %D is not compatible with DA local
> sizes %D %D\n") at err.c:482
> #5 0x0820d20a in DAVecGetArray (da=0x89f1870, vec=0x89f4770,
> array=0xbffcc770)
> at dagetarray.c:53
> #6 0x0804f162 in FormInitialGuess_physical (dmmg=0x898a560, X=0x899b070)
> at vecviewload_out.c:392
> #7 0x08052ced in DMMGSolve (dmmg=0x898a400) at damg.c:307
> #8 0x0804d479 in main (argc=Cannot access memory at address 0x348
> ) at vecviewload_out.c:186
>
> If I call
> // VecLoad(viewer,PETSC_NULL,&FIELD);
> VecLoadIntoVector(viewer,FIELD);
> [1]PETSC ERROR: VecLoadIntoVector() line 1016 in
> src/vec/vec/interface/vector.c Null Object: Parameter # 2
> [0]PETSC ERROR: VecLoadIntoVector() line 1016 in
> src/vec/vec/interface/vector.c Null Object: Parameter # 2
> [2]PETSC ERROR: VecLoadIntoVector() line 1016 in
> src/vec/vec/interface/vector.c Null Object: Parameter # 2
> [3]PETSC ERROR: VecLoadIntoVector() line 1016 in
> src/vec/vec/interface/vector.c Null Object: Parameter # 2
> Program received signal SIGABRT, Aborted.
> [Switching to Thread 0xb7c256b0 (LWP 1111)]
> 0xb7ef5410 in __kernel_vsyscall ()
> (gdb) where
> #0 0xb7ef5410 in __kernel_vsyscall ()
> #1 0xb7c75085 in raise () from /lib/tls/i686/cmov/libc.so.6
> #2 0xb7c76a01 in abort () from /lib/tls/i686/cmov/libc.so.6
> #3 0x08733fc9 in PetscAbortErrorHandler (line=1016,
> fun=0x885dc12 "VecLoadIntoVector", file=0x885d6b0 "vector.c",
> dir=0x885d6b9 "src/vec/vec/interface/", n=85, p=1,
> mess=0xbfc7f9d4 "Null Object: Parameter # 2", ctx=0x0) at errabort.c:62
> #4 0x086a8f3a in PetscError (line=1016, func=0x885dc12
> "VecLoadIntoVector",
> file=0x885d6b0 "vector.c", dir=0x885d6b9 "src/vec/vec/interface/", n=85,
> p=1, mess=0x885d6ed "Null Object: Parameter # %d") at err.c:482
> #5 0x085e985a in VecLoadIntoVector (viewer=0x8a08d10, vec=0x0)
> at vector.c:1016
> #6 0x0804f138 in FormInitialGuess_physical (dmmg=0x89a2880, X=0x89b34f0)
> at vecviewload_out.c:391
> #7 0x08052ce5 in DMMGSolve (dmmg=0x89a2720) at damg.c:307
> #8 0x0804d479 in main (argc=Cannot access memory at address 0x457
> ) at vecviewload_out.c:186
>
>
> So I am not sure how this binary write and read working?
>
> Did I miss sth?
>
> Thanks a lot!
>
> Rebecca
>
>
>
>
--
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
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20100310/5c7bb649/attachment.htm>
More information about the petsc-users
mailing list