Re: [petsc-dev] proposed minor PetscPartitioner changes

Thanks Barry, it works fine! :) Eric On 2022-08-31 16:28, Barry Smith wrote:   You  can use PetscObjectSetOptionsPrefix((PetscObject) partitioner,... On Aug 31, 2022, at 10:21 AM, Eric Chamberland wrote: Anyone have something to say about this? Thanks, Eric On 2022-07-06 22:32, Eric

Re: [petsc-dev] proposed minor PetscPartitioner changes

Anyone have something to say about this? Thanks, Eric On 2022-07-06 22:32, Eric Chamberland wrote: Hi, I was looking for PetscPartitionerSetOptionsPrefix function, but only found this post mentioning it. I was "blindly" coding with in mind the possibility to manage differe

Re: [petsc-dev] PETSc future starting as a new design layer that runs on top of PETSc 3?

easily and low-error prone call C, any other viable candidates? -- Eric Chamberland, ing., M. Ing Professionnel de recherche GIREF/Université Laval (418) 656-2131 poste 41 22 42

Re: [petsc-dev] proposed minor PetscPartitioner changes

changeable from options. Do you have anything against? Vaclav -- Eric Chamberland, ing., M. Ing Professionnel de recherche GIREF/Université Laval (418) 656-2131 poste 41 22 42

[petsc-dev] where is DATAFILESPATH ?

can I find the repository containing the data for these tests?... Or what is the right value for DATAFILESPATH? Thanks, Eric -- Eric Chamberland, ing., M. Ing Professionnel de recherche GIREF/Université Laval (418) 656-2131 poste 41 22 42

Re: [petsc-dev] Petsc "make test" have more failures for --with-openmp=1

aking news! :) Eric On 2021-03-30 1:47 p.m., Barry Smith wrote:   Eric,     How are things going on this OpenMP  front? Any bug fixes from hypre or SuperLU_DIST?     BTW: we have upgraded to OpenMPI 4.1 perhaps this resolves some issues?    Barry On Mar 22, 2021, at 2:07 PM, Eric C

Re: [petsc-dev] Petsc "make test" have more failures for --with-openmp=1

I added some information here: https://github.com/xiaoyeli/superlu_dist/issues/69#issuecomment-804318719 Maybe someone can say more than I on what PETSc tries to do with the 2 mentioned tutorials that are timing out... Thanks, Eric On 2021-03-15 11:31 a.m., Eric Chamberland wrote

Re: [petsc-dev] Petsc "make test" have more failures for --with-openmp=1

it.    Barry     I got the problem to occur with ex56 with 2 MPI ranks and 4 OpenMP threads, if I used less than 4 threads it did not generate an indefinite preconditioner. On Mar 14, 2021, at 1:18 PM, Eric Chamberland <mailto:eric.chamberl...@giref.ulaval.ca>> wrote: Don

Re: [petsc-dev] Petsc "make test" have more failures for --with-openmp=1

Hi again, ok, just saw that some matrices have lines of "0" in case of 3D hermite DOFs (ex: du/dz derivatives) when used into a 2D plane mesh... So, my last problem about hypre smoother is "normal". However, just to play with one of this matrix, I tried to do a "LU" with mumps icntl_24

Re: [petsc-dev] Petsc "make test" have more failures for --with-openmp=1

d not have seen the bug. Hopefully they can just fix it.    Barry     I got the problem to occur with ex56 with 2 MPI ranks and 4 OpenMP threads, if I used less than 4 threads it did not generate an indefinite preconditioner. On Mar 14, 2021, at 1:18 PM, Eric Chamberland <mailto:eric.chamberl.

Re: [petsc-dev] Petsc "make test" have more failures for --with-openmp=1

Reported timeout bugs to SuperLU_dist too: https://github.com/xiaoyeli/superlu_dist/issues/69 Eric On 2021-03-14 2:18 p.m., Eric Chamberland wrote: Done: https://github.com/hypre-space/hypre/issues/303 Maybe I will need some help about PETSc to answer their questions... Eric On 2021-03

Re: [petsc-dev] Petsc "make test" have more failures for --with-openmp=1

tps://github.com/hypre-space/hypre/issues> On Mar 14, 2021, at 3:44 AM, Eric Chamberland <mailto:eric.chamberl...@giref.ulaval.ca>> wrote: For us it clearly creates problems in real computations... I understand the need to have clean test for PETSc, but for me, it reveals that hypre isn't u

Re: [petsc-dev] Petsc "make test" have more failures for --with-openmp=1

suite for each test, you can add the extra flag in your command line TIMEOUT=180 (default is 60, units are seconds). Thanks, I’ll ping you on GitLab when I’ve got something ready for you to try, Pierre On 12 Mar 2021, at 8:54 PM, Eric Chamberland mailto:eric

Re: [petsc-dev] Petsc "make test" have more failures for --with-openmp=1

seconds). Noted, I will add this to my scripts and the dockerfiles too... Thanks, I’ll ping you on GitLab when I’ve got something ready for you to try, Thank you for all your work! Eric Pierre -- Eric Chamberland, ing., M. Ing Professionnel de recherche GIREF/Université Laval (418) 656-2131 poste 41 22 42

Re: [petsc-dev] Petsc "make test" have more failures for --with-openmp=1

the delay). Thanks, Eric On 2021-03-11 9:03 a.m., Eric Chamberland wrote: Hi Pierre, ok, that's interesting! I will try to build a docker image until tomorrow and give you the exact recipe to reproduce the bugs. Eric On 2021-03-11 2:46 a.m., Pierre Jolivet wrote: On 11 Mar 2021, at 6:16 AM, B

Re: [petsc-dev] Petsc "make test" have more failures for --with-openmp=1

tutorials-ex5f_superlu_dist_2.counts  ok ksp_ksp_tutorials-ex5f_superlu_dist_2 # SKIP Fortran required for this test    Barry On Mar 10, 2021, at 11:03 PM, Eric Chamberland <mailto:eric.chamberl...@giref.ulaval.ca>> wrote: Barry, to get a some follow up on --with-openmp=1 failures, shall I o

Re: [petsc-dev] Petsc "make test" have more failures for --with-openmp=1

Barry, to get a some follow up on --with-openmp=1 failures, shall I open gitlab issues for: a) all hypre failures giving DIVERGED_INDEFINITE_PC b) all superlu_dist failures giving different results with initia and "Exceeded timeout limit of 60 s" c) hpddm failures "free(): invalid next

Re: [petsc-dev] Petsc "make test" have more failures for --with-openmp=1

''],['mkl_blacs_openmpi'+ILP64+'']] mkl_blacs_32=[['mkl_blacs_intelmpi'],['mkl_blacs_mpich'],['mkl_blacs_sgimpt'],['mkl_blacs_openmpi']] On Mar 3, 2021, at 8:22 AM, Eric Chamberland <mailto:eric.chamberl...@giref.ulaval.ca>> wrote: Hi Pierre, On 2021-03-03 2:42 a.m., Pierre Jolivet wrot

Re: [petsc-dev] Petsc "make test" have more failures for --with-openmp=1

sts/counts/snes_tutorials-ex56_attach_mat_nearnullspace-0_bddc_approx_hypre.counts  ok snes_tutorials-ex56_attach_mat_nearnullspace-0_bddc_approx_hypre  ok diff-snes_tutorials-ex56_attach_mat_nearnullspace-0_bddc_approx_hypre         TEST arch-linux2-c-opt-ompi/tests/counts/ksp_ksp_tutorials-ex64_1.c

Re: [petsc-dev] Petsc "make test" have more failures for --with-openmp=1

Mar 2021, at 9:14 PM, Eric Chamberland <mailto:eric.chamberl...@giref.ulaval.ca>> wrote: Hi, It all started when I wanted to test PETSC/CUDA compatibility for our code. I had to activate --with-openmp to configure with --with-cuda=1 successfully. I then saw that PETSC_HAVE_OPENMP 

Re: [petsc-dev] Petsc "make test" have more failures for --with-openmp=1

linking configuration for example, should it be a good thing to have this (--with-openmp=1) tested into the pipelines (with external packages of course)? Thanks, Eric Thanks, Pierre On 2 Mar 2021, at 9:14 PM, Eric Chamberland <mailto:eric.chamberl...@giref.ulaval.ca>> wrote: Hi, It a

Re: [petsc-dev] Petsc "make test" have more failures for --with-openmp=1

    I have cc:ed Pierre so he can look at the HPDDM failures. On Mar 2, 2021, at 2:14 PM, Eric Chamberland <mailto:eric.chamberl...@giref.ulaval.ca>> wrote: Hi, It all started when I wanted to test PETSC/CUDA compatibility for our code. I had to activate --with-openmp to configu

Re: [petsc-dev] Petsc "make test" have more failures for --with-openmp=1

Pierre so he can look at the HPDDM failures. On Mar 2, 2021, at 2:14 PM, Eric Chamberland <mailto:eric.chamberl...@giref.ulaval.ca>> wrote: Hi, It all started when I wanted to test PETSC/CUDA compatibility for our code. I had to activate --with-openmp to configure with --wi

[petsc-dev] Petsc "make test" have more failures for --with-openmp=1

): https://giref.ulaval.ca/~cmpgiref/dernier_ompi/2021.03.02.02h16m01s_make_test.log https://giref.ulaval.ca/~cmpgiref/dernier_ompi/2021.03.02.02h16m01s_configure.log Thanks, Eric -- Eric Chamberland, ing., M. Ing Professionnel de recherche GIREF/Université Laval (418) 656-2131 poste 41 22 42

Re: [petsc-dev] Something changed with SBAIJ on april 17?

Perfect! I confirm the master branch is working again for me! Thanks a lot! :) Eric On 2019-04-23 10:48 p.m., Stefano Zampini wrote: Eric, Is in maint and master now On Apr 23, 2019, at 9:22 AM, Smith, Barry F. via petsc-dev wrote: Great, thanks Stefano for fixing them all On

[petsc-dev] mkl_cpardiso.c warnings

Hi, I just noticed there are warnings at compile time for some mkl_cpardiso.c routines: /pmi/cmpbib/compilation_BIB_gcc_redhat_petsc-master_debug/COMPILE_AUTO/petsc-master-debug/src/mat/impls/aij/mpi/mkl_cpardiso/mkl_cpardiso.c: In function ‘MatDestroy_MKL_CPARDISO’:

Re: [petsc-dev] Is mpich/master:a8a2b30fd21 tested with Petsc?

-clusters-and-hpc-technology/topic/797761 Eric On 19/04/18 09:01 AM, Eric Chamberland wrote: Hi, this morning, mpich/master with PETSc is 100% working again for us. Thanks to both commits: https://github.com/pmodels/mpich/commit/c597c8d79deea220a42751fda0f01ce70764c260 https://github.com

[petsc-dev] Valgrind report a bug in PetscOptionsFindPair_Private

Hi, since we recently upgraded our nighlty valgrind validations to PETSc 3.9.2, valgrind found a bug in PetscOptionsFindPair_Private, called from:

Re: [petsc-dev] Is mpich/master:a8a2b30fd21 tested with Petsc?

master branch soon. Thanks, Min On 2018/04/17 14:10, Eric Chamberland wrote: Hi, are we talking about the "tag" passed to MPI_Isend for example? but does that mean there is something to change for any MPI call which involves tags usage or is it only a PETSc "bad" tag us

Re: [petsc-dev] Is mpich/master:a8a2b30fd21 tested with Petsc?

y wrote: On Tue, 13 Mar 2018, Eric Chamberland wrote: Hi, each night we are testing mpich/master with our petsc-based code. I don't know if PETSc team is doing the same thing with mpich/master? (Maybe it is a good idea?) Everything was fine (except the issue https://github.com/pmodels/m

Re: [petsc-dev] [mpich-discuss] Is mpich/master:a8a2b30fd21 tested with Petsc?

ease. Thanks, Eric On 14/03/18 03:35 AM, Eric Chamberland wrote: Hi, fwiw, the actual mpich/master branch doesn't passes the PETSc "make test" after a fresh installation... It hangs just afer the 1 MPI process test, meaning it is locked into the 2 process test:

Re: [petsc-dev] [mpich-discuss] Is mpich/master:a8a2b30fd21 tested with Petsc?

s/ex19 run successfully with 1 MPI process ^Cmakefile:151: recipe for target 'test' failed make: [test] Interrupt (ignored) thanks, Eric On 13/03/18 08:07 AM, Eric Chamberland wrote: Hi, each night we are testing mpich/master with our petsc-based code. I don't know if PETSc team is doing the

[petsc-dev] Is mpich/master:a8a2b30fd21 tested with Petsc?

Hi, each night we are testing mpich/master with our petsc-based code.  I don't know if PETSc team is doing the same thing with mpich/master?   (Maybe it is a good idea?) Everything was fine (except the issue https://github.com/pmodels/mpich/issues/2892) up to commit 7b8d64debd, but since

[petsc-dev] petsc/master: superlu_dist now in PETSC_DIR/lib64

Hi, it is not a big issue, but I just wanted to know if it has been noticed that super-lu-dist 5.3.0 libs are now in PETSC_DIR/lib64 instead of PETSC_DIR/lib like all other libs? For me, after installation, it is the only package that have libs installed there...  I just added it to my

Re: [petsc-dev] petsc/master: unable to link in C++ with last night PETSC_WITH_EXTERNAL_LIB variable changes

On 12/02/18 12:11 PM, Jed Brown wrote: Eric Chamberland <eric.chamberl...@giref.ulaval.ca> writes: why does mpi stuff appears in the first one and not the second? I would prefer to have "-lmpi" in the second one too... When is it decided to "hide" exter

[petsc-dev] petsc/master: mkl_cpardiso.c broken

Hi, just to let know that since yesterday or before yesterday, I have errors in mkl_cpardiso.c on petsc/master: http://www.giref.ulaval.ca/~cmpgiref/petsc-master-debug/2018.01.31.02h00m01s_make.log other logs:

Re: [petsc-dev] ParMETIS change

if useful. Satish *From: *Eric Chamberland <mailto:eric.chamberl...@giref.ulaval.ca> *Sent: *Wednesday, September 13, 2017 8:41 AM *To: *petsc-dev <mailto:petsc-dev@mcs.anl.gov> *Subject: *[petsc-dev] ParMETIS change Hi, having trouble to compile our code with PETSc-master new default

[petsc-dev] ParMETIS change

Hi, having trouble to compile our code with PETSc-master new defaults options for PARMEtis. Parmetis was using double precision before... By curiosity, is there a big performance gain to use float? Thanks, Eric

Re: [petsc-dev] petscversion.h unchanged but API changed for MatGetSubMatrix and others

Hi, thanks for these infos. Isn't there a bug into the macros? Let's say I retreived a copy of petsc-master 3.7.2 then I have: #define PETSC_VERSION_RELEASE0 #define PETSC_VERSION_MAJOR 3 #define PETSC_VERSION_MINOR 7 #define PETSC_VERSION_SUBMINOR 2 Now an API changed in

[petsc-dev] petscversion.h unchanged but API changed for MatGetSubMatrix and others

Hi, Juste compiled with last night tarball of petsc-master and found that MatGetSubMatrix has changed to MatCreateSubMatrix. Fine, but I can't use the petscversion.h to switch into my code with #ifdef over any of: #define PETSC_VERSION_MAJOR 3 #define PETSC_VERSION_MINOR 7

Re: [petsc-dev] Is MatZeroRows too severe?

On 18/01/17 02:32 PM, Barry Smith wrote: Go for it. Ok, here it is! :) https://bitbucket.org/petsc/petsc/pull-requests/618/bug-fixe-the-debug_mode-for-abusive/diff It also includes an optimisation... Eric

Re: [petsc-dev] Is MatZeroRows too severe?

On 18/01/17 11:26 AM, Stefano Zampini wrote: You’re right. Maybe we should use MatFindZeroDiagonals if missing is true. I may try to do my first pull request if no one is under the way to correct this Is that ok? Thanks, Eric

Re: [petsc-dev] Is MatZeroRows too severe?

Hi Stefano, On 18/01/17 11:07 AM, Stefano Zampini wrote: Perhaps this is a valid fix - if (missing) SETERRQ1(PETSC_COMM_SELF,PETSC_ERR_ARG_WRONGSTATE,"Matrix is missing diagonal entry in row %D",d); + if (missing) { +for (i=0; i

[petsc-dev] Is MatZeroRows too severe?

Hi, we are calling MatZeroRows_SeqAIJ (petsc 3.7.5) with: N=3 rows[0]=15 rows[1]=30 rows[2]=65 diag = 1 x=0 b=0 and (Mat_SeqAIJ*)A->data->keepnonzeropattern is true. but MatMissingDiagonal_SeqAIJ finds a missing diagonal at line 1 and then return on error here: if (missing)

[petsc-dev] MatMissingDiagonal_MPIAIJ lack a reduce?

Hi, shouldn't there be a MPI_Reduce in MatMissingDiagonal_MPIAIJ so it is really collective as stated in documentation? Thanks, Eric

Re: [petsc-dev] Unable to access to MatSeqSBAIJGetArray_SeqSBAIJ from "outside" PETSc

Excellent! Thanks a lot! :) Eric Le 2017-01-06 à 19:37, Barry Smith a écrit : Added in master On Jan 6, 2017, at 2:46 PM, Eric Chamberland <eric.chamberl...@giref.ulaval.ca> wrote: Hi, we are trying to add a "petsc" developpement into our code but

Re: [petsc-dev] Unable to access to MatSeqSBAIJGetArray_SeqSBAIJ from "outside" PETSc

Thanks!!! :) Eric On 06/01/17 04:14 PM, Barry Smith wrote: Run ./configure with the additional option --with-visibility=0 We'll add a MatSeqSBAIJGetArray() Barry On Jan 6, 2017, at 2:46 PM, Eric Chamberland <eric.chamberl...@giref.ulaval.ca> wrote: Hi, we are trying

[petsc-dev] Unable to access to MatSeqSBAIJGetArray_SeqSBAIJ from "outside" PETSc

Hi, we are trying to add a "petsc" developpement into our code but we need to call MatSeqSBAIJGetArray_SeqSBAIJ for it. Unlike MatSeqAIJGetArray_SeqAIJ which is accessible with either MatSeqAIJGetArray or ierr =

Re: [petsc-dev] Cannot reuse matrix of type (null)

On 16/12/16 01:06 PM, Stefano Zampini wrote: What about raising an error from the interface if mat_reuse_matrix is requested with the same input and output matrix? Me too, I was asking myslef if this shall be done! Eric

Re: [petsc-dev] Cannot reuse matrix of type (null)

;type); } you are trying to call MatConvert with the reuse flag on a matrix which is not of type SeqDense. Can you provide a MWE to reproduce the problem? Thanks, Stefano 2016-12-16 16:39 GMT+03:00 Eric Chamberland <eric.chamberl...@giref.ulaval.ca>: Another precision: the problem may

Re: [petsc-dev] Cannot reuse matrix of type (null)

Ok, the offending call made is: ierr = (*conv)(mat,newtype,reuse,M);CHKERRQ(ierr); with: print newtype $1 = (MatType) 0x7fdc7bf8414a "seqdense" print reuse $2 = MAT_REUSE_MATRIX and mat is: (gdb) print *mat $4 = {hdr = {classid = 1211214, bops = {{getcomm = 0x7fdc6adda05b , view =

Re: [petsc-dev] Cannot reuse matrix of type (null)

Another precision: the problem may still be there since dec 15, not tonight... Eric On 16/12/16 08:30 AM, Eric Chamberland wrote: Hi, something seems broken with master branch since last night. With our nighlty tests, we now have many times this PETSC ERROR: [0]PETSC ERROR: Cannot reuse

[petsc-dev] Cannot reuse matrix of type (null)

Hi, something seems broken with master branch since last night. With our nighlty tests, we now have many times this PETSC ERROR: [0]PETSC ERROR: Cannot reuse matrix of type (null) [0]PETSC ERROR: See http://www.mcs.anl.gov/petsc/documentation/faq.html for trouble shooting. [0]PETSC ERROR:

[petsc-dev] Error in code documentation?

Hi, is there an error in code documentation in pvecimpl.h: typedef struct { VECHEADER PetscIntnghost; /* length of local portion including ghost padding */ isn't "nghost" the length of ghost padding only? Thanks, eric

Re: [petsc-dev] need C++ expert help

with m,n declared "int", for me 2 warnings arise from this line: first, "write" takes a size_t argument: write(fd,vals,(size_t)(m*n)*sizeof(double)) and returns a "ssize_t" value, so the rest of the line compiles fine like this: != (ssize_t)((size_t)(m*n)*sizeof(double) which gives:

[petsc-dev] Hijacked MPI calls...

Hi, I just discovered that all my calls to MPI are replaced by PETSc #define macros... I don't really care, but I now have this warning with clang 3.9.0: /pmi/cmpbib/compilation_BIB_gcc_redhat_squash/COMPILE_AUTO/GIREF/src/commun/Parallele/PABroadcast.icc:324:41: warning: possible misuse

Re: [petsc-dev] Are there both a function AND a #define to retrieve PETSc version?

On 28/08/16 06:24 PM, Jed Brown wrote: Eric Chamberland <eric.chamberl...@giref.ulaval.ca> writes: Le 2016-08-27 à 15:22, Jed Brown a écrit : Barry Smith <bsm...@mcs.anl.gov> writes: Well in Eric's case he knows he needs a bug fix in sub minor 3 and wants to gener

Re: [petsc-dev] Are there both a function AND a #define to retrieve PETSc version?

Le 2016-08-27 à 15:22, Jed Brown a écrit : Barry Smith writes: Well in Eric's case he knows he needs a bug fix in sub minor 3 and wants to generate an error if linked against an earlier one that does not have the bug fix. There needs to be a way for users to

Re: [petsc-dev] Are there both a function AND a #define to retrieve PETSc version?

Hi, Le 2016-08-27 à 15:07, Barry Smith a écrit : On Aug 27, 2016, at 1:59 PM, Jed Brown wrote: What about #define PetscInitialize(argc,argv,file,help)

[petsc-dev] Are there both a function AND a #define to retrieve PETSc version?

Hi, wanting to prevent a possible error: I compile and link (dynamically) my code with petsc-3.7.2 BUT I may change my LD_LIBRARY_PATH to point to 3.7.0... I would the like my code to stop and tell me I am using the wrong libs... I could add something like: if (PETSC_VERSION_MAJOR !=

[petsc-dev] petsc-master.tar.gz: unable to configure superlu_dist

Hi guys, something is broken for configuring superlu_dist with petsc-master.tar.gz available at 02h00 am: http://www.giref.ulaval.ca/~cmpgiref/petsc-master-debug/2016.03.22.02h00m01s_configure.log http://www.giref.ulaval.ca/~cmpgiref/petsc-master-debug/2016.03.22.02h00m01s_RDict.log I also

Re: [petsc-dev] PETSc release soon, request for input on needed fixes or enhancements

Hi Barry, the configuration of hypre looks broken since last night...? Here is the log issued from our automatic compilation of petsc-master.tar.gz at 2016-02-28 02h00: http://www.giref.ulaval.ca/~cmpgiref/petsc-master-debug/configure_20160228_0200.log Everything was fine until last

Re: [petsc-dev] petsc-master bug? "Too many pushes"

Le 2016-01-27 11:49, Barry Smith a écrit : Sorry about that. Fixed in master. Thanks for the quick fix!!! :) Eric

[petsc-dev] petsc-master bug? "Too many pushes"

Hi, since dec 15, we compile and test our code with the nightly "petsc-master" tarball (http://ftp.mcs.anl.gov/pub/petsc/petsc-master.tar.gz) and everything was fine until this morning. I noticed 4 tests in error: 0]PETSC ERROR: - Error Message

Re: [petsc-dev] Bug in MatShift_MPIAIJ ?

On 21/10/15 11:36 AM, Barry Smith wrote: On Oct 21, 2015, at 10:28 AM, Eric Chamberland <eric.chamberl...@giref.ulaval.ca> wrote: On 21/10/15 10:56 AM, Barry Smith wrote: I similarly saw the same behavior in the debugger that you reported and was mystified. This is why I

Re: [petsc-dev] Bug in MatShift_MPIAIJ ?

On 21/10/15 10:56 AM, Barry Smith wrote: I similarly saw the same behavior in the debugger that you reported and was mystified. This is why I called it a nasty bug. The setting of different nonew flag meant that one process called the Allreduce() at the end of MatAssemblyEnd_MPIAIJ()

Re: [petsc-dev] Bug in MatShift_MPIAIJ ?

SomewhatPetscValidLogicalCollectiveInt(__FILE__,comm); sorry, should be something like: ... SomewhatPetscValidLogicalCollectiveInt(function_that_sum_chars_to_int(__FILE__),comm); ... Eric

Re: [petsc-dev] Bug in MatShift_MPIAIJ ?

On 21/10/15 02:21 PM, Barry Smith wrote: Thanks. I don't think I need the Icount but. I think I can just generate an error if the FILE and LINE don't match exactly? Barry, I highly recommend you to use the static counter too otherwise you won't catch the following bug: if all processes

Re: [petsc-dev] Bug in MatShift_MPIAIJ ?

On 21/10/15 03:36 PM, Barry Smith wrote: Sure, but in that "case" the count would be identical on both processes anyways and so would not detect a problem? How could the count be different? hmmm, sorry, I think I am wrong and you are right. It won't stop the code earlier neither detect a

Re: [petsc-dev] Bug in MatShift_MPIAIJ ?

Thanks Barry! :) another question: while trying to understand this, I used the "-on_error_attach_debugger ddd" which worked, BUT the line where it was breaking all seemed ok to me. I mean, the value of "op" variable tested by at line petsc-3.6.2/src/mat/interface/matrix.c:5264 : 5256

Re: [petsc-dev] Bug in MatShift_MPIAIJ ?

Hi Barry, if I may suggest, is it possible to have this test case added as a nightly test since, I don't know but... a matrix with no local lines may be something not widely tested...??? Thanks! Eric On 20/10/15 10:47 PM, Barry Smith wrote: Eric, Thanks for the test case. I have

Re: [petsc-dev] Bug in MatShift_MPIAIJ ?

SET_2nd_PROC_TO_HAVE_NO_LOCAL_LINES to see that it runs fine if the 2nd process have local lines... The "magic" happens only if I call MatConvert on the "original" (C) matrix... Hope this helps! Thanks, Eric On 08/10/15 04:18 PM, Barry Smith wrote: On Oct 8, 20

Re: [petsc-dev] Bug in MatShift_MPIAIJ ?

Hi! hint: For petsc 3.6.1, if I apply the patches that Patrick gave (in the head of this thread), it all works correctly, Eric On 07/10/15 10:11 PM, Eric Chamberland wrote: Hi Jed, If I recall correctly, that patch is only checking for incorrect use by ensuring that the flags are set

Re: [petsc-dev] Bug in MatShift_MPIAIJ ?

I think the same problem is in MatShift_MPIBAIJ. I mean if I have a processor with no (local) lines at all in a matrix (which is my bugging case here), it should not call Mat*SetPreallocation at that point... That sounds ok to you? thanks, Eric On 08/10/15 02:45 PM, Eric Chamberland wrote

Re: [petsc-dev] Bug in MatShift_MPIAIJ ?

preallocated) { ierr = MatMPIAIJSetPreallocation(Y,1,NULL,0,NULL);CHKERRQ(ierr); } /*else if (!aij->nz) { ierr = MatSeqAIJSetPreallocation(maij->A,1,NULL);CHKERRQ(ierr); }*/ ierr = MatShift_Basic(Y,a);CHKERRQ(ierr); PetscFunctionReturn(0); } Eric On 08/10/15 09:45 AM, Eric Chamberland

Re: [petsc-dev] Bug in MatShift_MPIAIJ ?

On 08/10/15 03:30 PM, Satish Balay wrote: The commented code is sequential code - and shouldn't make a difference. ...but it does! Perhaps your application has other issues. Can you verify if your code is valgrind clean? http://www.mcs.anl.gov/petsc/documentation/faq.html#valgrind Yes

Re: [petsc-dev] Bug in MatShift_MPIAIJ ?

Hi Jed, If I recall correctly, that patch is only checking for incorrect use by ensuring that the flags are set collectively. If your code is correct then it doesn't affect you and if your code is incorrect, it will error now where as it would behave incorrectly before. The implementation was

Re: [petsc-dev] Bug in MatShift_MPIAIJ ?

Hi Barry, just compiled/tested with 3.6.2. On a 2 processes example, the non-debug version is hanging silently and indefinitely but in debug mode, it now abort when calling MatShift when a process has no entries: #0 0x7fffdbc6c065 in raise () from /lib64/libc.so.6 #1

Re: [petsc-dev] Bug in MatShift_MPIAIJ ?

/petsc-3.6.2_opt_openmpi-1.10_mkl_mt/lib/libpetsc.so.3.6(MatShift_MPIAIJ+0xf6) [0x7fa4eb90da45] #012: /opt/petsc-3.6.2_opt_openmpi-1.10_mkl_mt/lib/libpetsc.so.3.6(MatShift+0x9e) [0x7fa4eb6dbf94] while the other continued... thanks again! Eric On 07/10/15 03:02 PM, Eric Chamberland wrote: Hi

Re: [petsc-dev] PETSc release based on master before 9 pm tonight

On 06/08/2015 04:54 PM, Satish Balay wrote: Please do check src/docs/website/documentation/changes/dev.html and update it with any missing changes entries.. http://www.mcs.anl.gov/petsc/documentation/changes/dev.html thanks, Satish On Mon, 8 Jun 2015, Barry Smith wrote: If the nightly

Re: [petsc-dev] [petsc-users] Column #j is wrong in parallel from message Inserting a new nonzero (i, j) into matrix

On 03/25/2015 11:10 PM, Barry Smith wrote: On Mar 25, 2015, at 7:59 PM, Eric Chamberland eric.chamberl...@giref.ulaval.ca wrote: Hi Barry, On 03/25/2015 07:25 PM, Barry Smith wrote: This is an issue with any use of nested matrices (of which MPIBAIJ is a special, two level case). One

Re: [petsc-dev] [petsc-users] Column #j is wrong in parallel from message Inserting a new nonzero (i, j) into matrix

Hi Barry, On 03/25/2015 07:25 PM, Barry Smith wrote: This is an issue with any use of nested matrices (of which MPIBAIJ is a special, two level case). One way we could handle this is (I think universally) would be to introduce a unique PETSC_ERR_MAT_ENTRY_NONZERO error flag and then when

Re: [petsc-dev] installing external packages as shared?

On 09/08/2014 06:08 PM, Jed Brown wrote: Barry Smith bsm...@mcs.anl.gov writes: When should we install external packages as shared, currently if —with-PACKAGENAME-shared is set or if —with-shared-libraries is set we do it for some packages (like MPICH). Other packages, such as

Re: [petsc-dev] More prefetch

On 11/04/2014 04:40 PM, Barry Smith wrote: Unfortunately the test only checks if the link works, it doesn’t test if the prefetch actually does anything or might generate an error. Writing a proper (runtime) test for this, especially in a batch environment seems very difficult. Ok. #2- How

Re: [petsc-dev] More prefetch

Hi, it all looks like we have a bug with prefetch with PGI 14.7 + cpu specific compilation options. We have a signal 4 : illegal instruction (not reset when caught) within MatMult_SeqAIJ_Inode. The only thing we see there than can bug is the prefetch thing... We were asking ourself two

Re: [petsc-dev] Jed!!!

On 04/29/2014 03:14 PM, Jed Brown wrote: Eric Chamberland eric.chamberl...@giref.ulaval.ca writes: You're supposed to use MatGetLocalSubMatrix(). Translating from global indices to local indices is a disaster that we want to avoid. So we go the other way. Speak the language of split

Re: [petsc-dev] Jed!!!

On 04/27/2014 12:32 PM, Jed Brown wrote: Barry Smith bsm...@mcs.anl.gov writes: I am totally confused and likely users are to. Please point me to an example that actually uses Matnest the way it is “supposed to be used”. src/snes/examples/tutorials/ex28.c Note that this does not

Re: [petsc-dev] Fwd: MatSetValues with v=0

On 04/16/2014 06:00 PM, Barry Smith wrote: Removed in barry/rm-null-values-to-matsetvalues and next doh! ;-) Ok, I have done it with my own vector of zeros... Now I noticed another thing while measuring performances (notice: I used a PETSc compiled with debugging=yes... please don't

Re: [petsc-dev] Fwd: MatSetValues with v=0

On 04/17/2014 03:47 PM, Jed Brown wrote: Eric Chamberland eric.chamberl...@giref.ulaval.ca writes: Now I noticed another thing while measuring performances (notice: I used a PETSc compiled with debugging=yes... please don't flame me...): The first assembly for a 549252 square matrix takes me

Re: [petsc-dev] MatSetValues with v=0

On 04/12/2014 06:10 PM, Barry Smith wrote: Patrick, Currently it is supported for MatSetValues_SeqAIJ() and MatSetValues_MPIAIJ(). It is not handled for BAIJ and SBAIJ, the code while crash with segv in those cases (since we don’t check for a valid v pointer before use :-). But

Re: [petsc-dev] MatSetValues with v=0

On 04/16/2014 09:44 AM, Jed Brown wrote: Eric Chamberland eric.chamberl...@giref.ulaval.ca writes: I was writing (new) code which do the firsts MatSetValues after the MatXAIJSetPreallocation. We have to do this because the real non-zeros will be added later by a mix of ADD_VALUES

[petsc-dev] MatSetOption is logically collective even for MAT_ROW_ORIENTED.

Hi, On 10/04/2012 11:12 PM, Barry Smith wrote: supposedly in petsc-dev I just want to know if it will be included in next release... the problem still occurs with my freshly compiled 3-3.p5... Thanks! Eric

[petsc-dev] MatMatSolve_MUMPS

Hi, I just want to know if there is a known difficulty or unfeasible task if someone would like to implement MatMatSolve_MUMPS? Thanks, Eric

[petsc-dev] MatMatSolve_MUMPS

Jed, Hong, thank you for your answers! Eric Le 12/06/2012 01:25 PM, Jed Brown a ?crit : Eric, we would welcome an implementation that batched together a run-time-specified number of RHS vectors and used MUMPS' native multiple-RHS routines. The problem is that there has to be enough

[petsc-dev] OpenMPI petsc-2.3.3-p15

Hi, we have compile petsc-2.3.3-p15 with openMPI 1.3.4. To have it working with our C++ code, I did a #define OMPI_SKIP_MPICXX. In petsc.h, this is already done but for MPICH, on line #137: #define MPICH_SKIP_MPICXX 1 It would be nice to also have a: #define OMPI_SKIP_MPICXX 1 But I don't

PCApply_Shell ghosts...

Hi, we are using PETSc in our code and we have a problem with, I think, the ghosted values that we expect. We developed our own pre-conditioner, and try to use it in a parallel environment. With other precondtioners (PETSc built-in), everything works fine. But with our home-made one, here

Problem with VecPlaceArray_MPI VecResetArray

Hi, I try to use the VecPlaceArray/VecResetArray functions. Unfortunately, the call to VecPlaceArray only work the firsdt time I do it, even if I call VecResetArray before the next call... the code looks like this: VecPlaceArray(vec, myvec); ... VecResetArray(vec); VecPlaceArray(vec, myvec2);