| #
28f796ea
|
| 14-Jul-2020 |
Scott Kruger <kruger@txcorp.com> |
Enable HIP and SYCL configuration
This adds support for HIP and SYCL configuration.
HIP is both a compiler (hipcc) and a package to be searched for (hipblas, hipsparse) similar to CUDA. Language:
Enable HIP and SYCL configuration
This adds support for HIP and SYCL configuration.
HIP is both a compiler (hipcc) and a package to be searched for (hipblas, hipsparse) similar to CUDA. Language: HIP; Compiler: HIPCC (even though it is C++)
SYCL is currently implemented for searching only the dpcpp compiler. Because Intel's oneAPI is very similar to MKL, Blas/Lapack is handled by those packages. Language: SYCL; Compiler: SYCLCXX
Although it is a C++ compiler, both hipcc and dpcpp are very slow. Currently using the .hip.cpp and .sycl.cxx extensions although more work will be needed for testings. See the kokkos work.
This includes preliminary changes for pipeline testing The HIP_PLATFORM is an environment variable that chooses the backend. So one configuration file, but the platform is specified in .gitlab-ci.yml.
This includes beta versions of configuration files for what could be used on Frontier. This is both the HIP version and CUDA versions for enabling comparisons.
show more ...
|
| #
0d1e9396
|
| 22-Jun-2020 |
Satish Balay <balay@mcs.anl.gov> |
Merge branch 'barry/2020-06-13/fix-init-fptrap' into 'master'
Bug fix for initialization of _trapmode on Linux and MacOS when exception...
See merge request petsc/petsc!2877
|
| #
cc9df77e
|
| 14-Jun-2020 |
Barry Smith <bsmith@mcs.anl.gov> |
Bug fix for initialization of _trapmode on Linux and MacOS when exception handling of divide by zero is enabled
Cleanup and extension of floating point exception handling. Now properly turns
Bug fix for initialization of _trapmode on Linux and MacOS when exception handling of divide by zero is enabled
Cleanup and extension of floating point exception handling. Now properly turns off floating point trapping for Lapack when Fortran linker turns on divide by zero trapping
Commit-type: error-checking, bug-fix, portability-fix, feature, documentation Reported-by: Sanjay Govindjee <s_g@berkeley.edu>
show more ...
|
| #
6fe8eb35
|
| 09-Jun-2020 |
Satish Balay <balay@mcs.anl.gov> |
Merge branch 'MR--config-lowercase-pkgname' into 'master'
config: pkg-config: default to lowercase file.pc
See merge request petsc/petsc!2754
|
| #
2eefe1c6
|
| 07-Jun-2020 |
Jed Brown <jed@jedbrown.org> |
config: modernize DumpPkgconfig
|
| #
8244ab14
|
| 07-Jun-2020 |
Jed Brown <jed@jedbrown.org> |
pkg-config: switch to lowercase convention
Lowercase petsc.pc (versus PETSc.pc) is a more common convention. We'll keep the uppercase variant (on filesystems that support it) for backward compatibil
pkg-config: switch to lowercase convention
Lowercase petsc.pc (versus PETSc.pc) is a more common convention. We'll keep the uppercase variant (on filesystems that support it) for backward compatibility, but recommend using petsc.pc in new code.
show more ...
|
| #
3280a7ba
|
| 27-Apr-2020 |
Satish Balay <balay@mcs.anl.gov> |
Merge remote-tracking branch 'origin/maint'
|
| #
43171057
|
| 26-Apr-2020 |
Satish Balay <balay@mcs.anl.gov> |
Merge branch 'mg/lowercase-direct-h' into 'maint'
Fix detection of `direct.h` header file
See merge request petsc/petsc!2721
|
| #
4e8afd12
|
| 24-Apr-2020 |
Mosè Giordano <mose@gnu.org> |
Fix case of `windows.h` header
|
| #
7e4f0192
|
| 15-Apr-2020 |
Mosè Giordano <mose@gnu.org> |
Fix detection of `direct.h` header file
On case-sensitive filesystems the capitalisation of the file is important and the header file is correctly called `direct.h`, with lowercase `d`, see https://
Fix detection of `direct.h` header file
On case-sensitive filesystems the capitalisation of the file is important and the header file is correctly called `direct.h`, with lowercase `d`, see https://docs.microsoft.com/en-us/cpp/c-runtime-library/reference/mkdir-wmkdir?view=vs-2019#requirements
show more ...
|
| #
e08ecd42
|
| 05-Apr-2020 |
Satish Balay <balay@mcs.anl.gov> |
ci: fixes for vermin python syntax checker
vermin -vv --backport argparse -t=2.6- -t=3.4- config
|
| #
e15a6058
|
| 05-Apr-2020 |
Satish Balay <balay@mcs.anl.gov> |
Merge branch 'balay/ci-vermin/maint' into 'master'
Balay/ci vermin/maint
See merge request petsc/petsc!2686
|
| #
8f450857
|
| 05-Apr-2020 |
Satish Balay <balay@mcs.anl.gov> |
ci: fixes for vermin python syntax checker
vermin -vv --backport argparse -t=2.6- -t=3.4- config
|
| #
f1aa628f
|
| 18-Feb-2020 |
Satish Balay <balay@mcs.anl.gov> |
Merge remote-tracking branch 'origin/balay/ftn-updates'
Balay/ftn updates
See merge request petsc/petsc!2528
|
| #
6b541e4b
|
| 15-Feb-2020 |
Satish Balay <balay@mcs.anl.gov> |
Merge remote-tracking branch 'origin/maint'
|
| #
77764a62
|
| 15-Feb-2020 |
Satish Balay <balay@mcs.anl.gov> |
Merge branch 'jolivet/fix-mingw-toolchain' into 'maint'
Fix MinGW toolchain errors
See merge request petsc/petsc!2521
|
| #
ed50d614
|
| 15-Feb-2020 |
prj- <pierre.jolivet@enseeiht.fr> |
Fix MinGW toolchain errors
|
| #
257f4e5a
|
| 14-Feb-2020 |
Satish Balay <balay@mcs.anl.gov> |
more f90 fixes: remove PETSC_USING_F90 usage
|
| #
099d8087
|
| 14-Feb-2020 |
Satish Balay <balay@mcs.anl.gov> |
Merge remote-tracking branch 'origin/maint'
|
| #
81797890
|
| 14-Feb-2020 |
Satish Balay <balay@mcs.anl.gov> |
Merge remote-tracking branch 'origin/barry/2020-02-13/fix-separate-spelling' into maint
git grep -l "seperate" | xargs sed -i '' -e 's/seperate/separate/g'
See merge request petsc/petsc!2522
|
| #
8e5aa403
|
| 13-Feb-2020 |
Barry Smith <bsmith@mcs.anl.gov> |
git grep -l "seperate" | xargs sed -i '' -e 's/seperate/separate/g'
Commit-type: spelling-fix Reported-by: Pierre JOLIVET <Pierre.Jolivet@enseeiht.fr>
|
| #
cc59f4f8
|
| 08-Nov-2019 |
Satish Balay <balay@mcs.anl.gov> |
Merge remote-tracking branch 'origin/maint'
|
| #
56bc7b95
|
| 08-Nov-2019 |
Barry Smith <bsmith@mcs.anl.gov> |
Merge branch 'balay/fix-CXXPPFLAGS-usage' into 'maint'
build: fix usage of preprocessor flags across languages
See merge request petsc/petsc!2159
|
| #
0b119762
|
| 08-Oct-2019 |
Satish Balay <balay@mcs.anl.gov> |
build: fix usage of preprocessor flags across languages - consistently set flags CPP_FLAGS, FPP_FLAGS, CXXPP_FLAGS, PCPP_FLAGS - add PPPFLAGS, PFLAGS to map user-make-flags CFLAGS/CXXFLAGS or CPPFL
build: fix usage of preprocessor flags across languages - consistently set flags CPP_FLAGS, FPP_FLAGS, CXXPP_FLAGS, PCPP_FLAGS - add PPPFLAGS, PFLAGS to map user-make-flags CFLAGS/CXXFLAGS or CPPFLAGS/CXXPPFLAGS correctly to the PCC compile targets - use PCPP_FLAGS with PCC targets and CXXPP_FLAGS with CXX targets
Notes - CFLAGS/CPPFLAGS [API] for makefile does not override the the same variables specified to configure. [CFLAGS/CPPFAGS to configure are saved in CC_FLAGS, CPP_FLAGS] - For some reason CPPFLAGS gets added to both CC_FLAGS [via self.setCompilers.getCompilerFlags()] and CPP_FLAGS. Its not clear if one can be remoted - We do not have a clean C compile target like we have for Cxx - due to --with-clanguage [hence PCC etc intertwined with PETSC_COMPILE..]
show more ...
|
| #
80794ad0
|
| 24-Sep-2019 |
Satish Balay <balay@mcs.anl.gov> |
Merge branch 'barry/2019-07-27/error-check-mpiexec' into 'master'
Large number of fixes, optimizations for configure, speeds up the configure
See merge request petsc/petsc!2046
|