February 2012 - petsc-dev - mailman3.cels.anl.gov

Re: [petsc-dev] [petsc-users] Add a matrix to MatNest twice
by Barry Smith 21 Feb '12

21 Feb '12

On Feb 21, 2012, at 4:30 PM, Jed Brown wrote: > On Tue, Feb 21, 2012 at 15:37, Barry Smith <bsmith(a)mcs.anl.gov> wrote: > Why not just have MatAssemblyBegin_Nest() call the inner MatAssemblyBegin/End() together and stop the charade that there is any overlap of communication and computations etc anyway? > > (I pushed this.) > > So there is a very real latency issue in matrix assembly that comes from the reduction to determine how many receives are necessary. Due to MPI limitations, that code (PetscGatherNumberOfMessages() and PetscGatherMessageLengths()) is synchronizing, but MPI-3 will offer non-blocking collectives Oh, you mean using pthreads to spawn a thread that waits at blocking collective? :-) BTW; do you really think MPI-3 will exist? Or it should exist? Barry > that we could use for those operations. Now the two entrance points (MatAssemblyBegin() and MatAssemblyEnd()) are not sufficient to make progress on this task of assembly without also having an internal request system (where either a comm thread or callbacks from other library functions poked the progress along). > > There are also signs that sometime soon it will be common to have a comm thread that manages packing, in which case communication could actually start happening concurrently with computation. >

1 0

Re: [petsc-dev] Crash in MatDestroy_MUMPS
by Matthew Knepley 21 Feb '12

21 Feb '12

On Tue, Feb 21, 2012 at 3:55 PM, Jed Brown <jedbrown(a)mcs.anl.gov> wrote: > On Tue, Feb 21, 2012 at 15:38, John Fettig <john.fettig(a)gmail.com> wrote: > >> Is this likely memory corruption on my side or on the PETSc side, or >> something else entirely? I seem unable to reproduce it with an example, >> unfortunately. Running it through valgrind may take a week. > > > I don't see a reason why that specific line would naturally end up with an > invalid value, so I suspect memory corruption. Make a smaller problem so > that valgrind runs fast. > Put another way, even if this is our bug, it looks subtle and is unlikely to be found without valgrind. Thanks Matt -- 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

1 0

Re: [petsc-dev] Crash in MatDestroy_MUMPS
by Jed Brown 21 Feb '12

21 Feb '12

On Tue, Feb 21, 2012 at 15:38, John Fettig <john.fettig(a)gmail.com> wrote: > Is this likely memory corruption on my side or on the PETSc side, or > something else entirely? I seem unable to reproduce it with an example, > unfortunately. Running it through valgrind may take a week. I don't see a reason why that specific line would naturally end up with an invalid value, so I suspect memory corruption. Make a smaller problem so that valgrind runs fast.

1 0

Crash in MatDestroy_MUMPS
by John Fettig 21 Feb '12

21 Feb '12

Could you give me some pointers on how to debug this? Currently I am creating/destroying a KSP that uses MUMPS, and sometimes when it is destroyed I get a crash. It only happens in parallel. Here is the backtrace from petsc-dev for a crashed process: #0 0x00002b3a97f32c53 in PetscTrFreeDefault (aa=0x3ee0ec701867ea69, line=40, function=0x2b3a99509b34 "PetscComposedQuantitiesDestroy", file=0x2b3a99509888 "/home/jfe/local/petsc-dev/src/sys/objects/destroy.c", dir=0x2b3a995098bc "") at /home/jfe/local/petsc-dev/src/sys/memory/mtr.c:284 #1 0x00002b3a97f64b89 in PetscComposedQuantitiesDestroy (obj=0x1f090350) at /home/jfe/local/petsc-dev/src/sys/objects/destroy.c:40 #2 0x00002b3a980fc756 in VecScatterDestroy (ctx=0x19e6e938) at /home/jfe/local/petsc-dev/src/vec/vec/utils/vscat.c:1658 #3 0x00002b3a9884dd19 in MatDestroy_MUMPS (A=0x199a6cc0) at /home/jfe/local/petsc-dev/src/mat/impls/aij/mpi/mumps/mumps.c:484 #4 0x00002b3a981d376c in MatDestroy (A=0x14aaa5d0) at /home/jfe/local/petsc-dev/src/mat/interface/matrix.c:1046 #5 0x00002b3a98b158ba in PCReset_LU (pc=0x198cf370) at /home/jfe/local/petsc-dev/src/ksp/pc/impls/factor/lu/lu.c:174 #6 0x00002b3a98bfdd59 in PCReset (pc=0x198cf370) at /home/jfe/local/petsc-dev/src/ksp/pc/interface/precon.c:80 #7 0x00002b3a98cd0511 in KSPReset (ksp=0x14abb3f0) at /home/jfe/local/petsc-dev/src/ksp/ksp/interface/itfunc.c:718 #8 0x00002b3a98cd107d in KSPDestroy (ksp=0xd874a8) at /home/jfe/local/petsc-dev/src/ksp/ksp/interface/itfunc.c:755 I think a similar thing is happening in petsc-3.1-p8, here's a backtrace from there: #0 0x00002b20b2a964f8 in VMPI_Win_free () from /usr/local/encap/platform_mpi-8.01/lib/linux_amd64/libmpi.so.1 #1 0x00002b20b00d27ff in VecScatterDestroy_PtoP (ctx=0x1fe7bcf0) at vpscat.c:171 #2 0x00002b20b01018d5 in VecScatterDestroy (ctx=0x1fe7bcf0) at vscat.c:1638 #3 0x00002b20afd1249d in MatDestroy_MPIAIJ (mat=0x1cacb4d0) at mpiaij.c:909 #4 0x00002b20afc4c2cf in MatDestroy (A=0x1cacb4d0) at matrix.c:876 #5 0x00002b20afe972b0 in PCDestroy (pc=0x1032cf20) at precon.c:88 #6 0x00002b20afbbe78e in KSPDestroy (ksp=0x105f22d0) at itfunc.c:695 Is this likely memory corruption on my side or on the PETSc side, or something else entirely? I seem unable to reproduce it with an example, unfortunately. Running it through valgrind may take a week. Thanks, John

1 0

Re: [petsc-dev] Fwd: [petsc-maint #106588] ascii reader
by Barry Smith 21 Feb '12

21 Feb '12

The server for all the maints was hosed at ANL. It is supposedly fixed now. barry On Feb 21, 2012, at 12:45 PM, Matthew Knepley wrote: > I sent this yesterday! What is going on with petsc-main mail? > > Matt > > ---------- Forwarded message ---------- > From: Matthew Knepley <petsc-maint(a)mcs.anl.gov> > Date: Tue, Feb 21, 2012 at 12:44 PM > Subject: Re: [petsc-maint #106588] ascii reader > To: petsc-maint(a)mcs.anl.gov, jliu <jliu(a)ices.utexas.edu> > > > On Mon, Feb 20, 2012 at 10:05 PM, jliu <jliu(a)ices.utexas.edu> wrote: > > > The length seems to be wrong ... > > > > This is unhelpful. I can't understand how someone would be expected to > debug something > with the statement "this value seems to be wrong"? Here are things that > would be helpful > > a) What is the original vector size and ghost size? > > b) What is the size in the file? > > If you want this fixed, I would give us the information you would need to > understand it. > > Matt > > What do you mean by corrupt? The code I use to write the vector is: > > > > > > > > void PDNSolution::WriteBinary(const char * file_name) > > { > > PetscViewer viewer; > > PetscViewerCreate(PETSC_COMM_WORLD, &viewer); > > PetscViewerSetType(viewer, PETSCVIEWERBINARY); > > PetscViewerFileSetMode(viewer, FILE_MODE_WRITE); > > PetscViewerBinarySetMPIIO(viewer); > > PetscViewerBinarySkipInfo(viewer); > > PetscViewerFileSetName(viewer, file_name); > > > > VecView(solution, viewer); > > PetscViewerDestroy(&viewer); > > }; > > > > The input solution is a parallel vector with ghost component. I am not > > sure if that's enough for debugging. > > > > > > > > > > On 02/20/2012 09:59 PM, Matthew Knepley wrote: > > > On Mon, Feb 20, 2012 at 9:50 PM, jliu <jliu(a)ices.utexas.edu > > > <mailto:[email protected]>> wrote: > > > > > > > > > OK. > > > > > > This is my read code: > > > > > > 159 ierr = PetscViewerBinaryOpen(PETSC_COMM_SELF, > > > "Nonlinear_Diff_temp_900000000", > > > 160 FILE_MODE_READ, &view_in); > > > 161 CHKERRQ(ierr); > > > 162 > > > 163 ierr = VecLoad(sol, view_in); CHKERRQ(ierr); > > > > > > > > > The file now is clearly corrupt. So we can debug this if you make a > > > small example code > > > and send it. You can debug it with this code if you go into the > > > debugger and check the > > > size it reads in at the top of the file, and compare to the length of > > > the vector you wrote. > > > > > > Matt > > > > > > > > > > > > I get error message: > > > [0]PETSC ERROR: --------------------- Error Message > > > ------------------------------------ > > > [0]PETSC ERROR: Read from file failed! > > > [0]PETSC ERROR: Read past end of file! > > > [0]PETSC ERROR: > > > > > ------------------------------------------------------------------------ > > > [0]PETSC ERROR: Petsc Release Version 3.2.0, Patch 5, Sat Oct 29 > > > 13:45:54 CDT 2011 > > > [0]PETSC ERROR: See docs/changes/index.html for recent updates. > > > [0]PETSC ERROR: See docs/faq.html for hints about trouble shooting. > > > [0]PETSC ERROR: See docs/index.html for manual pages. > > > [0]PETSC ERROR: > > > > > ------------------------------------------------------------------------ > > > [0]PETSC ERROR: ./postprocess_error on a linux-gnu named hahn by jliu > > > Mon Feb 20 21:49:44 2012 > > > [0]PETSC ERROR: Libraries linked from > > > /org/groups/hughes/opt/lin/petsc-3.2-p5/linux-gnu-c-debug/lib > > > [0]PETSC ERROR: Configure run at Fri Feb 3 15:33:57 2012 > > > [0]PETSC ERROR: Configure options --with-cc=gcc --with-fc=gfortran > > > --download-f-blas-lapack=1 --download-mpich=1 > > > [0]PETSC ERROR: > > > > > ------------------------------------------------------------------------ > > > [0]PETSC ERROR: PetscBinaryRead() line 271 in > > > /org/groups/hughes/opt/lin/petsc-3.2-p5/src/sys/fileio/sysio.c > > > [0]PETSC ERROR: VecLoad_Binary() line 123 in > > > /org/groups/hughes/opt/lin/petsc-3.2-p5/src/vec/vec/utils/vecio.c > > > [0]PETSC ERROR: VecLoad_Default() line 359 in > > > /org/groups/hughes/opt/lin/petsc-3.2-p5/src/vec/vec/utils/vecio.c > > > [0]PETSC ERROR: VecLoad() line 1081 in > > > > > /org/groups/hughes/opt/lin/petsc-3.2-p5/src/vec/vec/interface/vector.c > > > [0]PETSC ERROR: User provided function() line 163 in > > > "unknowndirectory/"/h1/jliu/IsoPETSc/trunk/postprocess_error.cpp > > > > > > > > > > > > > > > > > > > > > On 02/20/2012 09:43 PM, Matthew Knepley wrote: > > > > On Mon, Feb 20, 2012 at 9:36 PM, jliu <jliu(a)ices.utexas.edu > > > <mailto:[email protected]> > > > > <mailto:[email protected] <mailto:[email protected]>>> > > wrote: > > > > > > > > On 02/20/2012 09:31 PM, Jed Brown wrote: > > > > > Always include petsc-maint. > > > > > > > > > > On Mon, Feb 20, 2012 at 21:28, jliu <jliu(a)ices.utexas.edu > > > <mailto:[email protected]> > > > > <mailto:[email protected] <mailto:[email protected]>> > > > > > <mailto:[email protected] <mailto:[email protected]> > > > <mailto:[email protected] <mailto:[email protected]>>>> wrote: > > > > > > > > > > I tried binary, but I always get an error: > > > > > > > > > > [0]PETSC ERROR: Read from file failed! > > > > > [0]PETSC ERROR: Read past end of file! > > > > > > > > > > I do not know what's wrong with the binary reader.. > > > > > > > > > > You have not explained what you wrote to the file or how you > > > > tried to > > > > > read the file. > > > > Oh, > > > > > > > > I write my binary file as: > > > > > > > > PetscViewer viewer; > > > > PetscViewerCreate(PETSC_COMM_WORLD, &viewer); > > > > PetscViewerSetType(viewer, PETSCVIEWERBINARY); > > > > PetscViewerFileSetMode(viewer, FILE_MODE_WRITE); > > > > PetscViewerBinarySetMPIIO(viewer); > > > > PetscViewerBinarySkipInfo(viewer); > > > > PetscViewerFileSetName(viewer, file_name); > > > > > > > > VecView(solution, viewer); > > > > PetscViewerDestroy(&viewer); > > > > > > > > > > > > As I said, check all return values so I can see the stack. How > > > hard is > > > > > > > > ierr = VecLoad(); CHKERRQ(ierr); > > > > > > > > Matt > > > > > > > > > > > > > > > > and I tried to read the saved file with: > > > > > > > > PetscViewerBinaryOpen(PETSC_COMM_SELF, > > > > "Nonlinear_Diff_temp_900000000", > > > > FILE_MODE_READ, &view_in); > > > > PetscViewerBinaryGetDescriptor(view_in,&fd); > > > > PetscBinaryRead(fd,&sz,1,PETSC_INT); > > > > cout<<sz<<endl; > > > > PetscBinaryRead(fd,sol,100,PETSC_SCALAR); > > > > VecLoad(sol, view_in); > > > > > > > > > > > > I vector saved is a parallel ghost vector. I don't know if > > > that is an > > > > issue. Else, I can not see what is the problem... > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > -- > > > > 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 > > > > > > > > > > > > > > > > > > -- > > > 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 > > > > > > > > > -- > 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 > > > > > -- > 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

1 0

Fwd: [petsc-maint #106588] ascii reader
by Matthew Knepley 21 Feb '12

21 Feb '12

I sent this yesterday! What is going on with petsc-main mail? Matt ---------- Forwarded message ---------- From: Matthew Knepley <petsc-maint(a)mcs.anl.gov> Date: Tue, Feb 21, 2012 at 12:44 PM Subject: Re: [petsc-maint #106588] ascii reader To: petsc-maint(a)mcs.anl.gov, jliu <jliu(a)ices.utexas.edu> On Mon, Feb 20, 2012 at 10:05 PM, jliu <jliu(a)ices.utexas.edu> wrote: > The length seems to be wrong ... > This is unhelpful. I can't understand how someone would be expected to debug something with the statement "this value seems to be wrong"? Here are things that would be helpful a) What is the original vector size and ghost size? b) What is the size in the file? If you want this fixed, I would give us the information you would need to understand it. Matt What do you mean by corrupt? The code I use to write the vector is: > > void PDNSolution::WriteBinary(const char * file_name) > { > PetscViewer viewer; > PetscViewerCreate(PETSC_COMM_WORLD, &viewer); > PetscViewerSetType(viewer, PETSCVIEWERBINARY); > PetscViewerFileSetMode(viewer, FILE_MODE_WRITE); > PetscViewerBinarySetMPIIO(viewer); > PetscViewerBinarySkipInfo(viewer); > PetscViewerFileSetName(viewer, file_name); > > VecView(solution, viewer); > PetscViewerDestroy(&viewer); > }; > > The input solution is a parallel vector with ghost component. I am not > sure if that's enough for debugging. > > > > > On 02/20/2012 09:59 PM, Matthew Knepley wrote: > > On Mon, Feb 20, 2012 at 9:50 PM, jliu <jliu(a)ices.utexas.edu > > <mailto:[email protected]>> wrote: > > > > > > OK. > > > > This is my read code: > > > > 159 ierr = PetscViewerBinaryOpen(PETSC_COMM_SELF, > > "Nonlinear_Diff_temp_900000000", > > 160 FILE_MODE_READ, &view_in); > > 161 CHKERRQ(ierr); > > 162 > > 163 ierr = VecLoad(sol, view_in); CHKERRQ(ierr); > > > > > > The file now is clearly corrupt. So we can debug this if you make a > > small example code > > and send it. You can debug it with this code if you go into the > > debugger and check the > > size it reads in at the top of the file, and compare to the length of > > the vector you wrote. > > > > Matt > > > > > > > > I get error message: > > [0]PETSC ERROR: --------------------- Error Message > > ------------------------------------ > > [0]PETSC ERROR: Read from file failed! > > [0]PETSC ERROR: Read past end of file! > > [0]PETSC ERROR: > > > ------------------------------------------------------------------------ > > [0]PETSC ERROR: Petsc Release Version 3.2.0, Patch 5, Sat Oct 29 > > 13:45:54 CDT 2011 > > [0]PETSC ERROR: See docs/changes/index.html for recent updates. > > [0]PETSC ERROR: See docs/faq.html for hints about trouble shooting. > > [0]PETSC ERROR: See docs/index.html for manual pages. > > [0]PETSC ERROR: > > > ------------------------------------------------------------------------ > > [0]PETSC ERROR: ./postprocess_error on a linux-gnu named hahn by jliu > > Mon Feb 20 21:49:44 2012 > > [0]PETSC ERROR: Libraries linked from > > /org/groups/hughes/opt/lin/petsc-3.2-p5/linux-gnu-c-debug/lib > > [0]PETSC ERROR: Configure run at Fri Feb 3 15:33:57 2012 > > [0]PETSC ERROR: Configure options --with-cc=gcc --with-fc=gfortran > > --download-f-blas-lapack=1 --download-mpich=1 > > [0]PETSC ERROR: > > > ------------------------------------------------------------------------ > > [0]PETSC ERROR: PetscBinaryRead() line 271 in > > /org/groups/hughes/opt/lin/petsc-3.2-p5/src/sys/fileio/sysio.c > > [0]PETSC ERROR: VecLoad_Binary() line 123 in > > /org/groups/hughes/opt/lin/petsc-3.2-p5/src/vec/vec/utils/vecio.c > > [0]PETSC ERROR: VecLoad_Default() line 359 in > > /org/groups/hughes/opt/lin/petsc-3.2-p5/src/vec/vec/utils/vecio.c > > [0]PETSC ERROR: VecLoad() line 1081 in > > > /org/groups/hughes/opt/lin/petsc-3.2-p5/src/vec/vec/interface/vector.c > > [0]PETSC ERROR: User provided function() line 163 in > > "unknowndirectory/"/h1/jliu/IsoPETSc/trunk/postprocess_error.cpp > > > > > > > > > > > > > > On 02/20/2012 09:43 PM, Matthew Knepley wrote: > > > On Mon, Feb 20, 2012 at 9:36 PM, jliu <jliu(a)ices.utexas.edu > > <mailto:[email protected]> > > > <mailto:[email protected] <mailto:[email protected]>>> > wrote: > > > > > > On 02/20/2012 09:31 PM, Jed Brown wrote: > > > > Always include petsc-maint. > > > > > > > > On Mon, Feb 20, 2012 at 21:28, jliu <jliu(a)ices.utexas.edu > > <mailto:[email protected]> > > > <mailto:[email protected] <mailto:[email protected]>> > > > > <mailto:[email protected] <mailto:[email protected]> > > <mailto:[email protected] <mailto:[email protected]>>>> wrote: > > > > > > > > I tried binary, but I always get an error: > > > > > > > > [0]PETSC ERROR: Read from file failed! > > > > [0]PETSC ERROR: Read past end of file! > > > > > > > > I do not know what's wrong with the binary reader.. > > > > > > > > You have not explained what you wrote to the file or how you > > > tried to > > > > read the file. > > > Oh, > > > > > > I write my binary file as: > > > > > > PetscViewer viewer; > > > PetscViewerCreate(PETSC_COMM_WORLD, &viewer); > > > PetscViewerSetType(viewer, PETSCVIEWERBINARY); > > > PetscViewerFileSetMode(viewer, FILE_MODE_WRITE); > > > PetscViewerBinarySetMPIIO(viewer); > > > PetscViewerBinarySkipInfo(viewer); > > > PetscViewerFileSetName(viewer, file_name); > > > > > > VecView(solution, viewer); > > > PetscViewerDestroy(&viewer); > > > > > > > > > As I said, check all return values so I can see the stack. How > > hard is > > > > > > ierr = VecLoad(); CHKERRQ(ierr); > > > > > > Matt > > > > > > > > > > > > and I tried to read the saved file with: > > > > > > PetscViewerBinaryOpen(PETSC_COMM_SELF, > > > "Nonlinear_Diff_temp_900000000", > > > FILE_MODE_READ, &view_in); > > > PetscViewerBinaryGetDescriptor(view_in,&fd); > > > PetscBinaryRead(fd,&sz,1,PETSC_INT); > > > cout<<sz<<endl; > > > PetscBinaryRead(fd,sol,100,PETSC_SCALAR); > > > VecLoad(sol, view_in); > > > > > > > > > I vector saved is a parallel ghost vector. I don't know if > > that is an > > > issue. Else, I can not see what is the problem... > > > > > > > > > > > > > > > > > > > > > > > > > > > -- > > > 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 > > > > > > > > > > > > -- > > 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 > > > -- 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 -- 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

1 0

Re: [petsc-dev] --download-mpich --download-openmpi and mpicc
by Sean Farley 21 Feb '12

21 Feb '12

> > I don't think this was the problem. We call updateCompilers() after the > install, so it should override anything the user puts in. > If I understand Barry correctly, it's that people who specify this: $ ./configure CC=mpicc --CXX=mpicxx --download-mpich clearly do not understand the --download-mpich option and should be told: ERROR: User needs to learn MPI; either use the cluster's custom built mpi XOR use --download-mpich because if configure && make succeed, then they will irrevocably use the mpiexec or mpicc in their $PATH and then use such great lines as "I have no idea what MPI is!" or "I just want to compile and run on my cluster! Why won't you let me do that?!"

1 0

Re: [petsc-dev] --download-mpich --download-openmpi and mpicc
by Matthew Knepley 21 Feb '12

21 Feb '12

On Tue, Feb 21, 2012 at 8:58 AM, Barry Smith <bsmith(a)mcs.anl.gov> wrote: > > Matt and Satish, > > --download-mpich and --download-openmpi should be immediately rejected > out of hand if the compilers start with mpi. Please add this support, it > is absurd we have to process email issues regarding this. I don't think this was the problem. We call updateCompilers() after the install, so it should override anything the user puts in. Matt > Barry > -- 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

1 0

Re: [petsc-dev] tests/benchmarks
by Matthew Knepley 21 Feb '12

21 Feb '12

On Tue, Feb 21, 2012 at 10:31 AM, Gerard Gorman <g.gorman(a)imperial.ac.uk>wrote: > Gerard Gorman emailed the following on 21/02/12 16:22: > > Hi > > > > I would like to harvest a selection of typical use cases for > > benchmarking PETSc/OpenMP. Ideally they would have features such as > > preallocating matrices rather than adding one at a time, be easy to > > configure for different problem sizes etc. Has anyone already put > > together such a list - what would be considered good/best practice here? > > > > Cheers > > Gerard > > > Further to this - do I recall someone saying they had unpushed bug fixes > to the test cases? For example, I've tried: > > ggorman@cynic:~/projects/petsc/petsc-dev/src/ksp/ksp/examples/tests$ > ./ex10 -m 4 > m = 2, N=375 > [0]PETSC ERROR: --------------------- Error Message > ------------------------------------ > [0]PETSC ERROR: Argument out of range! > [0]PETSC ERROR: New nonzero at (6,119) caused a malloc! > [0]PETSC ERROR: > ------------------------------------------------------------------------ > [0]PETSC ERROR: Petsc Development HG revision: > c7afeaa80089c71502abd0b0aca07e7ea2f3d431 HG Date: Tue Feb 21 09:48:53 > 2012 -0600 > [0]PETSC ERROR: See docs/changes/index.html for recent updates. > [0]PETSC ERROR: See docs/faq.html for hints about trouble shooting. > [0]PETSC ERROR: See docs/index.html for manual pages. > [0]PETSC ERROR: > ------------------------------------------------------------------------ > [0]PETSC ERROR: ./ex10 on a arch-linu named cynic by ggorman Tue Feb 21 > 16:28:33 2012 > [0]PETSC ERROR: Libraries linked from > /home/ggorman/projects/petsc/petsc-dev/arch-linux2-c-debug/lib > [0]PETSC ERROR: Configure run at Tue Feb 21 16:23:30 2012 > [0]PETSC ERROR: Configure options > [0]PETSC ERROR: > ------------------------------------------------------------------------ > [0]PETSC ERROR: MatSetValues_SeqAIJ() line 331 in > /home/ggorman/projects/petsc/petsc-dev/src/mat/impls/aij/seq/aij.c > [0]PETSC ERROR: MatSetValues() line 1142 in > /home/ggorman/projects/petsc/petsc-dev/src/mat/interface/matrix.c > [0]PETSC ERROR: AddElement() line 198 in src/ksp/ksp/examples/tests/ex10.c > [0]PETSC ERROR: GetElasticityMatrix() line 130 in > src/ksp/ksp/examples/tests/ex10.c > [0]PETSC ERROR: main() line 41 in src/ksp/ksp/examples/tests/ex10.c > -------------------------------------------------------------------------- > MPI_ABORT was invoked on rank 0 in communicator MPI_COMM_WORLD > with errorcode 63. > > NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes. > You may or may not see output from other processes, depending on > exactly when Open MPI kills them. > -------------------------------------------------------------------------- > > Pretty much the only value for the problem size (ie the -m option) is > the default. > Yes, the preallocation is not done right. We obviously do not use this one. Matt > Cheers > Gerard > > -- 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

1 0

Re: [petsc-dev] tests/benchmarks
by Gerard Gorman 21 Feb '12

21 Feb '12

Gerard Gorman emailed the following on 21/02/12 16:22: > Hi > > I would like to harvest a selection of typical use cases for > benchmarking PETSc/OpenMP. Ideally they would have features such as > preallocating matrices rather than adding one at a time, be easy to > configure for different problem sizes etc. Has anyone already put > together such a list - what would be considered good/best practice here? > > Cheers > Gerard > Further to this - do I recall someone saying they had unpushed bug fixes to the test cases? For example, I've tried: ggorman@cynic:~/projects/petsc/petsc-dev/src/ksp/ksp/examples/tests$ ./ex10 -m 4 m = 2, N=375 [0]PETSC ERROR: --------------------- Error Message ------------------------------------ [0]PETSC ERROR: Argument out of range! [0]PETSC ERROR: New nonzero at (6,119) caused a malloc! [0]PETSC ERROR: ------------------------------------------------------------------------ [0]PETSC ERROR: Petsc Development HG revision: c7afeaa80089c71502abd0b0aca07e7ea2f3d431 HG Date: Tue Feb 21 09:48:53 2012 -0600 [0]PETSC ERROR: See docs/changes/index.html for recent updates. [0]PETSC ERROR: See docs/faq.html for hints about trouble shooting. [0]PETSC ERROR: See docs/index.html for manual pages. [0]PETSC ERROR: ------------------------------------------------------------------------ [0]PETSC ERROR: ./ex10 on a arch-linu named cynic by ggorman Tue Feb 21 16:28:33 2012 [0]PETSC ERROR: Libraries linked from /home/ggorman/projects/petsc/petsc-dev/arch-linux2-c-debug/lib [0]PETSC ERROR: Configure run at Tue Feb 21 16:23:30 2012 [0]PETSC ERROR: Configure options [0]PETSC ERROR: ------------------------------------------------------------------------ [0]PETSC ERROR: MatSetValues_SeqAIJ() line 331 in /home/ggorman/projects/petsc/petsc-dev/src/mat/impls/aij/seq/aij.c [0]PETSC ERROR: MatSetValues() line 1142 in /home/ggorman/projects/petsc/petsc-dev/src/mat/interface/matrix.c [0]PETSC ERROR: AddElement() line 198 in src/ksp/ksp/examples/tests/ex10.c [0]PETSC ERROR: GetElasticityMatrix() line 130 in src/ksp/ksp/examples/tests/ex10.c [0]PETSC ERROR: main() line 41 in src/ksp/ksp/examples/tests/ex10.c -------------------------------------------------------------------------- MPI_ABORT was invoked on rank 0 in communicator MPI_COMM_WORLD with errorcode 63. NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes. You may or may not see output from other processes, depending on exactly when Open MPI kills them. -------------------------------------------------------------------------- Pretty much the only value for the problem size (ie the -m option) is the default. Cheers Gerard

1 0