[clblas] branch master updated (27ab572 -> d16f7b3)
Ghislain Vaillant
ghisvail-guest at moszumanska.debian.org
Thu Jan 14 20:07:38 UTC 2016
This is an automated email from the git hooks/post-receive script.
ghisvail-guest pushed a change to branch master
in repository clblas.
from 27ab572 Merge pull request #163 from hughperkins/fix-teardown
new c6d971f bump develop branch version number to 2.9.0
new d987b0a Update CMakeLists.txt
new 902ccec Merge pull request #157 from TimmyLiu/develop
new 3681c78 Fix catch-22 in build order, following pull-163, where init.c compile fails because AutoGemmClKernels.h hasnt been built yet
new 0fc3d3f Merge pull request #170 from hughperkins/fix-buggette-in-pull-163
new 45ba4b3 merge master branch to develop branch. please only make pull requests to develop branch
new 75b0f92 Merge pull request #174 from TimmyLiu/develop
new 56f7957 fix hard-coding of opencl version to 2.0; fix 1d initialization of 2d arrays.
new 2ea8c8a Fix teardown of UserGemmClKernels
new 484dc16 Merge pull request #175 from hughperkins/fix-teardown-userkernels-rebase-develop
new 16b0f9e add missing include on stdlib.h
new 57e25b9 Merge pull request #171 from ghisvail/fix/missing-stdlib-include
new dff63b4 Fix kernel crash on nvidia, caused by float4 alignemtn error, see https://github.com/clMathLibraries/clBLAS/issues/108 for more details
new a50aaef Merge pull request #115 from hughperkins/develop
new f54bc62 release event obj in sample code
new f888be0 Merge pull request #178 from TimmyLiu/develop
new c96ebc6 find libblas.so correctly. use libblas.so for testing by default instead of libacml.so
new 26b40b8 need gfortran for testing with netlib blas
new 7495ed6 try to get gfortran working on travis ci
new 2d67b2e remove version number
new 09e7100 Merge pull request #181 from TimmyLiu/develop
new 89f62e5 Update gtest link to point to github
new 0725afa Merge pull request #183 from hughperkins/update-gtest-link
new 767f2a3 Fixed clang build error error: embedding a #pragma directive within macro arguments is not supported
new 57e1e65 Merge pull request #190 from notorca/master
new 99b3931 adding AutoGemm kernel selection logic for Fiji
new 1e7aec0 Merge pull request #194 from guacamoleo/develop
new 0d2fb3a if BUILD_TEST is not set fortran compiler should not be required.
new 5bfe00d Merge pull request #193 from TimmyLiu/develop
new 4914167 add perf data for fiji
new c2e7334 Merge pull request #198 from tingxingdong/performanceData1
new 843bff5 Add caching mechanism based on context and device for gemm and trsm
new 596925c Work around for an nvidia bug when querying kernel function name
new f7397fe Add cl_khr_fp64 when using double precision
new 1983aaf Changing AutoGemm scripts to work with both python2 and python3
new bef2f6b Attempt to build from souce if build from binary fails
new ad00a99 Merge pull request #199 from arrayfire/arrayfire-release-test
new 7aba3d0 Find the python interpreter using cmake
new 590b47d Merge pull request #200 from arrayfire/arrayfire-release-test
new 8851fc1 attempts to fix travis ci build with missing fglrx
new 22da020 previous commit failed to verify khrons' certificate. try dist trusty now
new d340d57 typo fix
new d43c42b not build client for now
new fffd478 bump master branch version number to 2.10.0
new d16f7b3 Merge pull request #210 from TimmyLiu/master
The 45 revisions listed above as "new" are entirely new to this
repository and will be described in separate emails. The revisions
listed as "adds" were already present in the repository and have only
been added to this reference.
Summary of changes:
.gitignore | 3 +
.travis.yml | 49 ++-
.../clblas290_fijinano_cgemm_col_nt_1520.csv} | 361 ++++++++--------
.../clblas290_fijinano_dgemm_col_nt_1520.csv} | 361 ++++++++--------
.../clblas290_fijinano_sgemm_col_nt_1520.csv} | 362 ++++++++--------
.../clblas290_fijinano_zgemm_col_nt_1520.csv} | 339 +++++++--------
src/CMakeLists.txt | 10 +-
src/FindNetlib.cmake | 5 +-
src/library/CMakeLists.txt | 11 +-
src/library/blas/AutoGemm/AutoGemm.py | 8 +-
src/library/blas/AutoGemm/AutoGemmParameters.py | 70 ++-
src/library/blas/AutoGemm/AutoGemmTeardown.h | 10 +
src/library/blas/AutoGemm/Includes.py | 4 +-
src/library/blas/AutoGemm/KernelOpenCL.py | 30 +-
src/library/blas/AutoGemm/KernelParameters.py | 42 +-
src/library/blas/AutoGemm/KernelSelection.py | 30 +-
src/library/blas/AutoGemm/KernelsToPreCompile.py | 7 +-
.../UserGemmKernelSources/UserGemmClKernels.cc | 54 +++
.../UserGemmKernelSources/UserGemmClKernels.h | 22 +-
.../UserGemmKernelSourceIncludes.h | 4 +-
.../create_user_gemm_cl_kernels.py | 44 ++
.../dgemm_Col_NN_B0_MX048_NX048_KX08_src.cpp | 3 +-
.../dgemm_Col_NN_B1_MX048_NX048_KX08_src.cpp | 3 +-
.../dgemm_Col_NT_B0_MX048_NX048_KX08_src.cpp | 3 +-
.../dgemm_Col_NT_B1_MX048_NX048_KX08_src.cpp | 3 +-
.../dgemm_Col_TN_B0_MX048_NX048_KX08_src.cpp | 3 +-
.../dgemm_Col_TN_B1_MX048_NX048_KX08_src.cpp | 3 +-
.../sgemm_Col_NN_B0_MX032_NX032_KX16_src.cpp | 2 +-
.../sgemm_Col_NN_B0_MX064_NX064_KX16_src.cpp | 2 +-
.../sgemm_Col_NN_B0_MX096_NX096_KX16_src.cpp | 2 +-
...sgemm_Col_NN_B1_MX032_NX032_KX16_BRANCH_src.cpp | 2 +-
.../sgemm_Col_NN_B1_MX032_NX032_KX16_src.cpp | 2 +-
.../sgemm_Col_NN_B1_MX064_NX064_KX16_src.cpp | 2 +-
.../sgemm_Col_NN_B1_MX096_NX096_KX16_src.cpp | 2 +-
.../sgemm_Col_NT_B0_MX032_NX032_KX16_src.cpp | 2 +-
.../sgemm_Col_NT_B0_MX064_NX064_KX16_src.cpp | 2 +-
.../sgemm_Col_NT_B0_MX096_NX096_KX16_src.cpp | 2 +-
...sgemm_Col_NT_B1_MX032_NX032_KX16_BRANCH_src.cpp | 2 +-
.../sgemm_Col_NT_B1_MX032_NX032_KX16_src.cpp | 2 +-
.../sgemm_Col_NT_B1_MX064_NX064_KX16_src.cpp | 2 +-
.../sgemm_Col_NT_B1_MX096_NX096_KX16_src.cpp | 2 +-
.../sgemm_Col_TN_B0_MX032_NX032_KX16_src.cpp | 2 +-
.../sgemm_Col_TN_B0_MX064_NX064_KX16_src.cpp | 2 +-
.../sgemm_Col_TN_B0_MX096_NX096_KX16_src.cpp | 2 +-
...sgemm_Col_TN_B1_MX032_NX032_KX16_BRANCH_src.cpp | 2 +-
.../sgemm_Col_TN_B1_MX032_NX032_KX16_src.cpp | 2 +-
.../sgemm_Col_TN_B1_MX064_NX064_KX16_src.cpp | 2 +-
.../sgemm_Col_TN_B1_MX096_NX096_KX16_src.cpp | 2 +-
src/library/blas/gens/clTemplates/ger.cl | 6 +-
src/library/blas/init.c | 4 +-
.../blas/trtri/diag_dtrtri_lower_128_16.cpp | 6 +-
.../blas/trtri/diag_dtrtri_upper_128_16.cpp | 6 +-
.../blas/trtri/diag_dtrtri_upper_192_12.cpp | 6 +-
src/library/blas/xgemm.cc | 162 ++++---
src/library/blas/xtrsm.cc | 480 ++++++++++++---------
src/samples/example_chbmv.c | 3 +
src/samples/example_chemm.cpp | 5 +-
src/samples/example_cher.c | 2 +
src/samples/example_cher2k.c | 3 +
src/samples/example_cherk.cpp | 3 +
src/samples/example_chpmv.c | 3 +
src/samples/example_chpr.c | 2 +
src/samples/example_csscal.c | 5 +-
src/samples/example_dtrmv.c | 3 +
src/samples/example_isamax.c | 3 +
src/samples/example_sasum.c | 3 +
src/samples/example_saxpy.c | 3 +
src/samples/example_scopy.c | 3 +
src/samples/example_sdot.c | 3 +
src/samples/example_sgbmv.c | 3 +
src/samples/example_sgemm.c | 3 +
src/samples/example_sgemv.c | 3 +
src/samples/example_sger.c | 3 +
src/samples/example_snrm2.c | 3 +
src/samples/example_srot.c | 4 +-
src/samples/example_srotg.c | 3 +
src/samples/example_srotm.c | 3 +
src/samples/example_srotmg.c | 3 +
src/samples/example_ssbmv.c | 3 +
src/samples/example_sscal.c | 3 +
src/samples/example_sspmv.c | 3 +
src/samples/example_sspr.c | 2 +
src/samples/example_sspr2.c | 2 +
src/samples/example_sswap.c | 3 +
src/samples/example_ssymm.c | 3 +
src/samples/example_ssymv.c | 3 +
src/samples/example_ssyr.c | 2 +
src/samples/example_ssyr2.c | 2 +
src/samples/example_ssyr2k.c | 2 +
src/samples/example_ssyrk.c | 3 +
src/samples/example_stbmv.c | 2 +
src/samples/example_stbsv.c | 3 +
src/samples/example_stpmv.c | 2 +
src/samples/example_stpsv.c | 3 +
src/samples/example_strmm.c | 3 +
src/samples/example_strmv.c | 2 +
src/samples/example_strsm.c | 2 +
src/samples/example_strsv.c | 3 +
src/samples/example_zhemv.cpp | 3 +
src/samples/example_zher2.c | 2 +
src/samples/example_zhpr2.c | 2 +
src/tests/gtest.cmake | 4 +-
102 files changed, 1554 insertions(+), 1136 deletions(-)
copy doc/performance/{clBLAS_2.6.0/W9100/peak_dp.csv => clBLAS_2.9.0/FIJINANO/clblas290_fijinano_cgemm_col_nt_1520.csv} (51%)
copy doc/performance/{clBLAS_2.6.0/W9100/peak_sp.csv => clBLAS_2.9.0/FIJINANO/clblas290_fijinano_dgemm_col_nt_1520.csv} (52%)
copy doc/performance/{clBLAS_2.6.0/W9100/peak_dp.csv => clBLAS_2.9.0/FIJINANO/clblas290_fijinano_sgemm_col_nt_1520.csv} (51%)
copy doc/performance/{clBLAS_2.6.0/S9150/peak_sp.csv => clBLAS_2.9.0/FIJINANO/clblas290_fijinano_zgemm_col_nt_1520.csv} (51%)
create mode 100644 src/library/blas/AutoGemm/AutoGemmTeardown.h
create mode 100644 src/library/blas/AutoGemm/UserGemmKernelSources/UserGemmClKernels.cc
create mode 100644 src/library/blas/AutoGemm/UserGemmKernelSources/create_user_gemm_cl_kernels.py
--
Alioth's /usr/local/bin/git-commit-notice on /srv/git.debian.org/git/debian-science/packages/clblas.git
More information about the debian-science-commits
mailing list