Typos
Improve -help output.* remove redundant separators* separators more clearly separate sections* improve help for -log_exclude* separate PetscOptionsGetViewer() help section* get rid of PetscOpti
Improve -help output.* remove redundant separators* separators more clearly separate sections* improve help for -log_exclude* separate PetscOptionsGetViewer() help section* get rid of PetscOptionsCheckInitial_Components()
show more ...
Variable/function renaming for easy reading
Add Fortran support for PetscSubcommCommit-type: fortran, exampleReported-by: PAEZ ESPEJO Miguel-Angel EUROGICIEL INGENIERIE <Miguel-Angel.PAEZESPEJO@cea.fr>
Remove usage of PETSC_HAVE_FORTRAN_STDCALL, PETSC_BLASLAPACK_STDCALL, HAVE_FORTRAN_MIXED_STR_ARG flags - as Compaq f90 compiler is no longer supported
Add support for users to pass in FFLAGS the option for promoting integers to 8 bytes.Tested for ifort and gfortranTested with MPICH and OpenMPI and MPI-UniSince MPI does not provide any mechani
Add support for users to pass in FFLAGS the option for promoting integers to 8 bytes.Tested for ifort and gfortranTested with MPICH and OpenMPI and MPI-UniSince MPI does not provide any mechanism for integer promotion the code very crudely adjusts thevalues in the mpif.h to mark them as 4 byte integers this is not neccessary portable for all MPI implementationsThis is an extremely fragile option that we should probably not be supporting.Users must insure their Fortran codes passes only 4 byte integers to MPI callsPETSc Fortran examples were all fixed to explicitly use 4 byte integers for MPI calls.Commit-type: portability-fix, testing-fix, featureTime: 20 hoursReported-by: Dmitry Melnichuk <dmitry.melnichuk@geosteertech.com>
fix bug with Fortran PetscInit
Merge branch 'balay/update-solaris-mpich' [PR #1856]* balay/update-solaris-mpich:mpich: update opensolaris from using mpich-3.3a2 to mpich-3.3.1
mpich: update opensolaris builds of PETSc from using mpich-3.3a2 to mpich-3.3.1The MPICH 3.3.1 HWLOC component with Solaris compilers and optimization -O hangs while calling MPI.The fix suggested
mpich: update opensolaris builds of PETSc from using mpich-3.3a2 to mpich-3.3.1The MPICH 3.3.1 HWLOC component with Solaris compilers and optimization -O hangs while calling MPI.The fix suggested is to set HWLOC_COMPONENTS="-x86" in the environmentThis is handled in PETSc by1) setting the HWLOC_COMPONENTS="-x86" in configure before testing the MPI to prevent hangs during configure1a) having configure print a warning that setting this environmental variable may be needed to prevent hangs in running code2) having PetscInitialize() set HWLOC_COMPONENTS="-x86" so the user does not have to, which handles most cases except those where the user directly calls MPI_Init() and for that we have the warning at configure time (can't have the warning at run time because the user calls MPI_Init() before they get to PETSc code.Requested-by: Junchao Zhang <jczhang@mcs.anl.gov>Thanks-to: Satish Balay <balay@mcs.anl.gov>Thanks-to: Jed Brown <jed@jedbrown.org>
Remove additional uneeded values from petscconf.h; reorganize handling of default debugger, make it lldb on darwinCommit-type: style-fixThanks-to: Jed Brown <jed@jedbrown.org>
Merged in barry/cuda-multigrid-test (pull request #1763)Various improvements for GPUs (mostly for performance and CUDA)
Non-numeric optimizations focused on AIJ, MatFDColoring, and DMCreateMatrix_DA_*AIJ, looking to improve performance in GPU environments1) PetscCalloc*() now uses system calloc()2) Merged some Pets
Non-numeric optimizations focused on AIJ, MatFDColoring, and DMCreateMatrix_DA_*AIJ, looking to improve performance in GPU environments1) PetscCalloc*() now uses system calloc()2) Merged some PetscMalloc*()3) Eliminated unneeded PetscCalloc*()4) Removed some memory allocations and copies in MatFDColoringSetUp(), added local variables for better compiler optimization5) Added MatSetValues_SeqAIJ_SortedFull(), added MatSetOption(MAT_SORTED_FULL)6) Optimized DMCreateMatrix_DA_*AIJ for nonperiodic case to automatically have sorted columns (faster MatSetValues() times)7) Eliminated call to PetscMemzero() in PetscFree()Commit-type: style-fix, feature
Added PetscCUDAInitialize() which the user may call in the body of the program, -cuda_initialize 0 prevents PetscInitialize() from initializing CUDAAdd -cuda_initialize 0 as options database for CU
Added PetscCUDAInitialize() which the user may call in the body of the program, -cuda_initialize 0 prevents PetscInitialize() from initializing CUDAAdd -cuda_initialize 0 as options database for CUDA tests so machines won't load cuBLAS unless actuallyutilizing it on the machine, update examples to filter option-cuda_initializeReported-by: Karl Rupp <me@karlrupp.net>Commit-type: feature
Man pages: remove .keywords: fieldsThis field is recognized by doctext from Sowing (http://wgropp.cs.illinois.edu/projects/software/sowing/doctext/node20.htm#Node20),but it doesn't seem to be unif
Man pages: remove .keywords: fieldsThis field is recognized by doctext from Sowing (http://wgropp.cs.illinois.edu/projects/software/sowing/doctext/node20.htm#Node20),but it doesn't seem to be uniformly used or maintained.Thus, remove all .keywords: fields, and a following blank line, if present.This is accomplished with GNU sed (gsed on OS X), with the following commands.*Warning* that this type of command can corrupt a .git directory,so be cautious in reusing or modifying these commands. They first lookfor and delete matching lines with a following line consisting of only whitespace,and then delete any remaining matching lines. find src/ -type f -exec gsed -i '/keywords:/ {N; /\n\s*$/d}' {} + find src/ -type f -exec gsed -i '/keywords:/d' {} + find include/ -type f -exec gsed -i '/keywords:/ {N; /\n\s*$/d}' {} + find include/ -type f -exec gsed -i '/keywords:/d' {} +Hints on the sed command obtained from: https://unix.stackexchange.com/questions/100754/how-to-delete-a-specific-line-and-the-following-blank-line-using-gnu-sed
MPIU_SIZE_T: define at run-time
MPIU_FORTRANADDR: define at run-time
MPI_2INT is not always a macro. And its part of MPI-1 anyway. So assume is it always exists and works.This issue comes up with HPE MPTReported-by: "Kokron, Daniel S. (ARC-606.2)[InuTeq, LLC]" <da
MPI_2INT is not always a macro. And its part of MPI-1 anyway. So assume is it always exists and works.This issue comes up with HPE MPTReported-by: "Kokron, Daniel S. (ARC-606.2)[InuTeq, LLC]" <daniel.s.kokron@nasa.gov
Add if (*ierr) return checks for all Fortran stubs that utilize FREECHAR()Otherwise the ierr is overwritten by the FREECHAR() thus returning a zero error code when it should notCommit-type: bug-f
Add if (*ierr) return checks for all Fortran stubs that utilize FREECHAR()Otherwise the ierr is overwritten by the FREECHAR() thus returning a zero error code when it should notCommit-type: bug-fixReported-by: Tim Steinhoff <kandanovian@gmail.com>
Merge remote-tracking branch 'origin/jczhang/feature-mumps-omp'
Remove CHKFORTRANNULLOBJECTDEREFERENCE since it is no longer needed with Fortran null equal to C nullCommit-type: feature
Revise PETSc shared memory communicator keyval interfaceNames like Petsc_ShmComm_keyval etc are conforming to existing keyvals like Petsc_InnerComm_keyval
Merge remote-tracking branch 'origin/barry/adios'
adios: fixes- fix configure to pass in MPI compilers [instead of --with-mpi=DIR. Similarly attempt to use PETSc configure detected values for hdf5, netcdf, zlib etc..- fix compiler error with usin
adios: fixes- fix configure to pass in MPI compilers [instead of --with-mpi=DIR. Similarly attempt to use PETSc configure detected values for hdf5, netcdf, zlib etc..- fix compiler error with using '0' for ADIOS_STATISTICS_FLAG- declare Petsc_adios_group as PETSC_EXTERN - this way its useable from -lpetscvec [with-visibility=1]- also move Petsc_adios_group to petscimpl.h - to avoid duplicate definitions- fix fortran examples [by adding adios initialization to fortran interface- fix compile error with ex10
Moved the rest of the PETSc common block variables over to petscsys moduleCommit-type: feature
Move PETSC_COMM_WORLD and PETSC_COMM_SELF from common blocks to petscsys moduleReason: Build on Apple with seperate libraries resulted in duplicate copies of the symbols (set to null).Eventually
Move PETSC_COMM_WORLD and PETSC_COMM_SELF from common blocks to petscsys moduleReason: Build on Apple with seperate libraries resulted in duplicate copies of the symbols (set to null).Eventually all the common block entries should be moved to the petscsys.h module. This is a test for next to find any problems.Commit-type: portability-fix
12345678910