<div dir="ltr">On Thu, Aug 15, 2013 at 6:51 PM, Satish Balay <span dir="ltr"><<a href="mailto:balay@mcs.anl.gov" target="_blank">balay@mcs.anl.gov</a>></span> wrote:<br><div class="gmail_extra"><div class="gmail_quote">
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Alternative way to handle this is:<br>
<br>
>>><br>
#define HAVE_CUBLAS<br>
#include "magma.h"<br>
<<<<br>
<br>
both in configure test - and petsc code that uses magma.h<br>
[but I'm not sure if this is possible with configure/package.py]<br></blockquote><div><br></div><div>This is even worse. Just let Barry pollute our code with bad solutions that we will eventually have</div><div>to fix up by doing it right.</div>
<div><br></div><div>I think we need to impose some order on things we incorporate, if only to reduce the maintenance</div><div>burden on us.</div><div><br></div><div> Matt</div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Satish<br>
<br>
<br>
On Thu, 15 Aug 2013, Barry Smith wrote:<br>
<br>
><br>
> Damn Matt, why do you persist in giving such bad idealist advice. Magma is badly designed so be it, we can't do anything about that. You can't suddenly impose that magma do something reasonable. We have to live with their silly decisions.<br>
><br>
> Harshad should use insertCompilerFlag() to simply add this flag -DHAVE_CUBLAS so that it gets used by everything compiled by BuildSystem. Unfortunately since the compiler stuff is handled in such an arcane way in BuildSystem I don't know on what object this method should be called on. self? self.framework? some compiler beasty? Anyways just tell him how call insertCompilerFlag() and the proper place to call it from within magma.py and everything will be fine.<br>
><br>
> Barry<br>
<div class="HOEnZb"><div class="h5">><br>
><br>
><br>
> On Aug 15, 2013, at 3:38 PM, Matthew Knepley <<a href="mailto:knepley@gmail.com">knepley@gmail.com</a>> wrote:<br>
><br>
> > On Thu, Aug 15, 2013 at 3:35 PM, Harshad Sahasrabudhe <<a href="mailto:hsahasra@purdue.edu">hsahasra@purdue.edu</a>> wrote:<br>
> > I don't know how to write a config header. Can you give me an example package which I can look at?<br>
> ><br>
> > ---magmaconfig.h<br>
> ><br>
> > #define HAVE_CUBLAS 1<br>
> ><br>
> ><br>
> > Matt<br>
> ><br>
> > ----- Original Message -----<br>
> > From: "Matthew Knepley" <<a href="mailto:knepley@gmail.com">knepley@gmail.com</a>><br>
> > To: "Harshad Sahasrabudhe" <<a href="mailto:hsahasra@purdue.edu">hsahasra@purdue.edu</a>><br>
> > Cc: "Karl Rupp" <<a href="mailto:rupp@mcs.anl.gov">rupp@mcs.anl.gov</a>>, "For users of the development version of PETSc" <<a href="mailto:petsc-dev@mcs.anl.gov">petsc-dev@mcs.anl.gov</a>><br>
> > Sent: Thursday, August 15, 2013 4:29:12 PM<br>
> > Subject: Re: [petsc-dev] Error while building MAGMA from PETSc<br>
> ><br>
> ><br>
> > On Thu, Aug 15, 2013 at 3:24 PM, Harshad Sahasrabudhe < <a href="mailto:hsahasra@purdue.edu">hsahasra@purdue.edu</a> > wrote:<br>
> ><br>
> ><br>
> ><br>
> ><br>
> > It is getting passed... Here is one line from configure.log:<br>
> ><br>
> > nvcc -g -DADD_ -Xcompiler -fno-strict-aliasing -DGPUSHMEM=130 -DHAVE_CUBLAS -gencode arch=compute_13,code=sm_13 -gencode arch=compute_13,code=compute_13 -gencode arch=compute_10,code=compute_10 -I/home/harshad/cuda-5.5/include -I../include -I../control -c <a href="http://zgeadd_batched.cu" target="_blank">zgeadd_batched.cu</a> -o zgeadd_batched.cu_o<br>
> ><br>
> ><br>
> > During the Magma build. HOWEVER, when you compile with this header, you need to have this defined as well.<br>
> > I suggest writing a config header for that package since this "define on the command<br>
> > line stuff" is horrible. Then include that header in the regular Magma header that you are checking for.<br>
> ><br>
> ><br>
> > Matt<br>
> ><br>
> ><br>
> ><br>
> > ----- Original Message -----<br>
> > From: "Karl Rupp" < <a href="mailto:rupp@mcs.anl.gov">rupp@mcs.anl.gov</a> ><br>
> > To: "Harshad Sahasrabudhe" < <a href="mailto:hsahasra@purdue.edu">hsahasra@purdue.edu</a> ><br>
> > Cc: "Barry Smith" < <a href="mailto:bsmith@mcs.anl.gov">bsmith@mcs.anl.gov</a> >, "For users of the development version of PETSc" < <a href="mailto:petsc-dev@mcs.anl.gov">petsc-dev@mcs.anl.gov</a> ><br>
> ><br>
> ><br>
> > Sent: Thursday, August 15, 2013 4:23:37 PM<br>
> > Subject: Re: [petsc-dev] Error while building MAGMA from PETSc<br>
> ><br>
> > Hi again,<br>
> ><br>
> > > Yes, -DHAVE_CUBLAS is defined while compilation, but somehow I'm<br>
> > getting this error.<br>
> ><br>
> > Make sure this also gets passed to the CUDA compiler (nvcc).<br>
> ><br>
> > Best regards,<br>
> > Karli<br>
> ><br>
> ><br>
> > ><br>
> > > ----- Original Message -----<br>
> > > From: "Karl Rupp" < <a href="mailto:rupp@mcs.anl.gov">rupp@mcs.anl.gov</a> ><br>
> > > To: "Harshad Sahasrabudhe" < <a href="mailto:hsahasra@purdue.edu">hsahasra@purdue.edu</a> ><br>
> > > Cc: "Barry Smith" < <a href="mailto:bsmith@mcs.anl.gov">bsmith@mcs.anl.gov</a> >, "For users of the development version of PETSc" < <a href="mailto:petsc-dev@mcs.anl.gov">petsc-dev@mcs.anl.gov</a> ><br>
> > > Sent: Thursday, August 15, 2013 4:18:57 PM<br>
> > > Subject: Re: [petsc-dev] Error while building MAGMA from PETSc<br>
> > ><br>
> > > Hey,<br>
> > ><br>
> > > apparently there must be one out of {CUBLAS, clAmdBlas, ICC for MIC}<br>
> > > installed. CUBLAS is already made available via --with-cuda, so you<br>
> > > should be fine pointing Magma to the CUDA folder. We don't have package<br>
> > > checks for clAmdBlas or MIC.<br>
> > ><br>
> > > Best regards,<br>
> > > Karli<br>
> > ><br>
> > ><br>
> > > On 08/15/2013 03:13 PM, Harshad Sahasrabudhe wrote:<br>
> > >> How does the testing work? What headers and functions do I need to include in magma.py?<br>
> > >><br>
> > >> ----- Original Message -----<br>
> > >> From: "Barry Smith" < <a href="mailto:bsmith@mcs.anl.gov">bsmith@mcs.anl.gov</a> ><br>
> > >> To: "Harshad Sahasrabudhe" < <a href="mailto:hsahasra@purdue.edu">hsahasra@purdue.edu</a> ><br>
> > >> Cc: "Jed Brown" < <a href="mailto:jedbrown@mcs.anl.gov">jedbrown@mcs.anl.gov</a> >, "For users of the development version of PETSc" < <a href="mailto:petsc-dev@mcs.anl.gov">petsc-dev@mcs.anl.gov</a> ><br>
> > >> Sent: Thursday, August 15, 2013 4:01:29 PM<br>
> > >> Subject: Re: [petsc-dev] Error while building MAGMA from PETSc<br>
> > >><br>
> > >><br>
> > >> Possible ERROR while running preprocessor: In file included from /home/harshad/workspace/petsc-hg/arch-linux2-c-debug/include/magmablas_z.h:14:0,<br>
> > >> from /home/harshad/workspace/petsc-hg/arch-linux2-c-debug/include/magmablas.h:12,<br>
> > >> from /home/harshad/workspace/petsc-hg/arch-linux2-c-debug/include/magma.h:15,<br>
> > >> from /tmp/petsc-Z_wD6e/config.headers/conftest.c:3:<br>
> > >> /home/harshad/workspace/petsc-hg/arch-linux2-c-debug/include/magma_types.h:146:6: error: #error "One of HAVE_CUBLAS, HAVE_clAmdBlas, or HAVE_MIC must be defined. This typically happens in Makefile.internal."<br>
> > >> ret = 256<br>
> > >> error message = {In file included from /home/harshad/workspace/petsc-hg/arch-linux2-c-debug/include/magmablas_z.h:14:0,<br>
> > >> from /home/harshad/workspace/petsc-hg/arch-linux2-c-debug/include/magmablas.h:12,<br>
> > >> from /home/harshad/workspace/petsc-hg/arch-linux2-c-debug/include/magma.h:15,<br>
> > >> from /tmp/petsc-Z_wD6e/config.headers/conftest.c:3:<br>
> > >> /home/harshad/workspace/petsc-hg/arch-linux2-c-debug/include/magma_types.h:146:6: error: #error "One of HAVE_CUBLAS, HAVE_clAmdBlas, or HAVE_MIC must be defined. This typically happens in Makefile.internal."<br>
> > >><br>
> > >><br>
> > >> On Aug 15, 2013, at 2:43 PM, Harshad Sahasrabudhe < <a href="mailto:hsahasra@purdue.edu">hsahasra@purdue.edu</a> > wrote:<br>
> > >><br>
> > >>> Attached is the diff and configure.log<br>
> > >>><br>
> > >>> ----- Original Message -----<br>
> > >>> From: "Jed Brown" < <a href="mailto:jedbrown@mcs.anl.gov">jedbrown@mcs.anl.gov</a> ><br>
> > >>> To: "Matthew Knepley" < <a href="mailto:knepley@gmail.com">knepley@gmail.com</a> ><br>
> > >>> Cc: "Harshad Sahasrabudhe" < <a href="mailto:hsahasra@purdue.edu">hsahasra@purdue.edu</a> >, "For users of the development version of PETSc" < <a href="mailto:petsc-dev@mcs.anl.gov">petsc-dev@mcs.anl.gov</a> ><br>
> > >>> Sent: Thursday, August 15, 2013 3:33:33 PM<br>
> > >>> Subject: Re: [petsc-dev] Error while building MAGMA from PETSc<br>
> > >>><br>
> > >>><br>
> > >>><br>
> > >>> ... and send the changes you have made. One way to do that is to 'git add config/PETSc/packages/magma.py' and send the output if 'git diff master'.<br>
> > >>> On Aug 15, 2013 2:28 PM, "Matthew Knepley" < <a href="mailto:knepley@gmail.com">knepley@gmail.com</a> > wrote:<br>
> > >>><br>
> > >>><br>
> > >>><br>
> > >>> On Thu, Aug 15, 2013 at 1:56 PM, Harshad Sahasrabudhe < <a href="mailto:hsahasra@purdue.edu">hsahasra@purdue.edu</a> > wrote:<br>
> > >>><br>
> > >>><br>
> > >>><br>
> > >>><br>
> > >>> Hi,<br>
> > >>><br>
> > >>> Now MAGMA compiles correctly and libmagma.a is generated, however PETSc says<br>
> > >>><br>
> > >>> Downloaded magma could not be used. Please check install in /home/harshad/workspace/petsc-hg/arch-linux2-c-debug<br>
> > >>><br>
> > >>> I am copying the generated library and include files to the install directory using<br>
> > >>><br>
> > >>> includeDir = os.path.join(self.installDir, self.includedir)<br>
> > >>> output,err,ret = PETSc.package.NewPackage.executeShellCommand('cd '+self.packageDir+' && mv -f lib/*.* '+libDir+'/. && cp -f include/*.* '+includeDir+'/.', timeout=2500, log = self.framework.log)<br>
> > >>><br>
> > >>> I need help in resolving this error.<br>
> > >>><br>
> > >>><br>
> > >>><br>
> > >>> Always send configure.log.<br>
> > >>><br>
> > >>><br>
> > >>> Matt<br>
> > >>><br>
> > >>><br>
> > >>> Thanks,<br>
> > >>> Harshad<br>
> > >>><br>
> > >>> ----- Original Message -----<br>
> > >>> From: "Matthew Knepley" < <a href="mailto:knepley@gmail.com">knepley@gmail.com</a> ><br>
> > >>> To: "Harshad Sahasrabudhe" < <a href="mailto:hsahasra@purdue.edu">hsahasra@purdue.edu</a> ><br>
> > >>> Cc: "For users of the development version of PETSc" < <a href="mailto:petsc-dev@mcs.anl.gov">petsc-dev@mcs.anl.gov</a> ><br>
> > >>> Sent: Wednesday, August 14, 2013 6:27:04 PM<br>
> > >>> Subject: Re: [petsc-dev] Error while building MAGMA from PETSc<br>
> > >>><br>
> > >>><br>
> > >>> On Wed, Aug 14, 2013 at 5:23 PM, Harshad Sahasrabudhe < <a href="mailto:hsahasra@purdue.edu">hsahasra@purdue.edu</a> > wrote:<br>
> > >>><br>
> > >>><br>
> > >>><br>
> > >>><br>
> > >>><br>
> > >>><br>
> > >>> Hi,<br>
> > >>><br>
> > >>> I'm trying the write the package compiler for MAGMA using PETSc build system. I'm getting the following error while building MAGMA. With the error, I'm also getting a couple hundred warnings:<br>
> > >>><br>
> > >>> sormqr_gpu_2stage.cpp: In function ‘magma_int_t magma_sormqr_gpu_2stages(char, char, magma_int_t, magma_int_t, magma_int_t, float*, magma_int_t, float*, magma_int_t, float*, magma_int_t, magma_int_t*)’:<br>
> > >>> sormqr_gpu_2stage.cpp:98: warning: unused variable ‘c_one’<br>
> > >>> sormqr_gpu_2stage.cpp:107: warning: unused variable ‘lwkopt’<br>
> > >>> strsm_m.cpp: In function ‘magma_int_t magma_strsm_m(magma_int_t, char, char, char, char, magma_int_t, magma_int_t, float, float*, magma_int_t, float*, magma_int_t)’:<br>
> > >>> strsm_m.cpp:181: warning: unused variable ‘jj’<br>
> > >>> strsm_m.cpp:181: warning: unused variable ‘jjb’<br>
> > >>> sgeev.cpp: In function ‘magma_int_t magma_sgeev(char, char, magma_int_t, float*, magma_int_t, float*, float*, float*, magma_int_t, float*, magma_int_t, float*, magma_int_t, magma_int_t*)’:<br>
> > >>> sgeev.cpp:126: warning: unused variable ‘c_n1’<br>
> > >>> nvcc fatal : redefinition of argument 'gpu-architecture'<br>
> > >>><br>
> > >>><br>
> > >>> What is the easiest way to debug this kind of an error?<br>
> > >>><br>
> > >>><br>
> > >>><br>
> > >>> 1) Make the MAGMA build system print out the full compile line<br>
> > >>><br>
> > >>><br>
> > >>> 2) Make the MAGMA people clean up their sloppy code, which would eliminate the warnings<br>
> > >>><br>
> > >>><br>
> > >>> Matt<br>
> > >>><br>
> > >>><br>
> > >>><br>
> > >>><br>
> > >>> Thanks,<br>
> > >>> Harshad<br>
> > >>><br>
> > >>><br>
> > >>><br>
> > >>><br>
> > >>><br>
> > >>> --<br>
> > >>> What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead.<br>
> > >>> -- Norbert Wiener<br>
> > >>><br>
> > >>><br>
> > >>><br>
> > >>><br>
> > >>> --<br>
> > >>> What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead.<br>
> > >>> -- Norbert Wiener<diff_magma><configure.log><br>
> > >><br>
> > ><br>
> ><br>
> ><br>
> ><br>
> ><br>
> ><br>
> > --<br>
> > What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead.<br>
> > -- Norbert Wiener<br>
> ><br>
> ><br>
> ><br>
> > --<br>
> > What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead.<br>
> > -- Norbert Wiener<br>
><br>
><br>
</div></div></blockquote></div><br><br clear="all"><div><br></div>-- <br>What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead.<br>
-- Norbert Wiener
</div></div>