Commit Graph

  • f5243e8e1f
    Add compiler option to avx512 test and hide test output Martin Kroeker 2018-06-30 23:47:44 +0200
  • ba8388cee0
    Merge pull request #1651 from martin-frbg/avx512-nodgemm Martin Kroeker 2018-06-30 17:48:03 +0200
  • 6e54b0a027
    Disable the 16x2 DTRMM kernel on SkylakeX as well Martin Kroeker 2018-06-30 17:31:06 +0200
  • 40c8cbc3bf
    Merge pull request #1650 from martin-frbg/avx512-nodgemm Martin Kroeker 2018-06-30 13:05:46 +0200
  • d3c9eb4c7d
    Merge pull request #1639 from martin-frbg/dyn_list Martin Kroeker 2018-06-30 13:05:30 +0200
  • f0a8dc2eec
    Disable the AVX512 DGEMM kernel for now Martin Kroeker 2018-06-30 11:34:48 +0200
  • cc92257ea6
    Update Makefile Martin Kroeker 2018-06-27 00:09:21 +0200
  • 2aba1b1658
    Merge branch 'develop' into nofort Martin Kroeker 2018-06-27 00:07:32 +0200
  • 8396e9e777
    Handle NOFORTRAN=0 Martin Kroeker 2018-06-27 00:00:27 +0200
  • bfad307ed7
    Merge pull request #1647 from martin-frbg/armv7-dot Martin Kroeker 2018-06-26 22:27:30 +0200
  • b83e4c60c7
    Remove premature exit for INC_X or INC_Y zero Martin Kroeker 2018-06-26 20:46:42 +0200
  • e344db269b
    Remove premature exit for INC_X or INC_Y zero Martin Kroeker 2018-06-26 20:45:57 +0200
  • 545b82efd3
    Remove premature exit for INC_X or INC_Y zero Martin Kroeker 2018-06-26 20:45:00 +0200
  • e322a951fe
    Remove premature exit for INC_X or INC_Y zero Martin Kroeker 2018-06-26 20:44:13 +0200
  • ff2f171036
    Merge pull request #1644 from martin-frbg/revert-filterout Martin Kroeker 2018-06-26 10:15:15 +0200
  • 092175cfec
    Revert changes to NOFORTRAN handling from 952541e Martin Kroeker 2018-06-26 08:09:52 +0200
  • 750162a05f
    Try gradual fallback for cores not in the dynamic core list Martin Kroeker 2018-06-25 21:02:31 +0200
  • e6d93f20f1
    Merge pull request #2 from martin-frbg/develop Martin Kroeker 2018-06-25 20:48:10 +0200
  • c38c65eb65
    Merge pull request #1 from xianyi/develop Martin Kroeker 2018-06-25 20:45:56 +0200
  • ce3651516f
    Merge pull request #1642 from oon3m0oo/develop Martin Kroeker 2018-06-25 19:23:40 +0200
  • 0144068537 Rewrite &= -> = and simplify the initial blocking phase. Craig Donner 2018-06-25 13:53:11 +0100
  • 1833a67071
    Add support for a user-defined list of dynamic targets Martin Kroeker 2018-06-23 19:42:15 +0200
  • 0b2b83d9ed
    Add support for a user-defined list of dynamic targets Martin Kroeker 2018-06-23 19:41:32 +0200
  • 62cf769aa6
    Merge pull request #1638 from martin-frbg/issue1637 Martin Kroeker 2018-06-23 15:01:02 +0200
  • eb71d61c7c
    Expose CBLAS interface to BLAS extensions iXamin Martin Kroeker 2018-06-23 13:31:09 +0200
  • 9cf22b7d91
    Build cblas_iXamin interfaces Martin Kroeker 2018-06-23 13:27:30 +0200
  • cc66743b66
    Merge pull request #1634 from oon3m0oo/develop Martin Kroeker 2018-06-21 21:01:03 +0200
  • 2aa0a5804e Use BLAS rather than CBLAS in test_fork.c (#1626) oon3m0oo 2018-06-21 17:47:45 +0100
  • 28c28ed275 Fix data races reported by TSAN. Craig Donner 2018-06-21 11:13:57 +0100
  • f4f1107325 Use BLAS rather than CBLAS in test_fork.c Craig Donner 2018-06-19 16:31:26 +0100
  • a399d00425 Further improvements to memory.c. (#1625) oon3m0oo 2018-06-20 21:04:03 +0100
  • f66b9c8826
    Merge pull request #1630 from martin-frbg/x86-march Martin Kroeker 2018-06-20 21:51:57 +0200
  • 2946c46024
    Merge pull request #1631 from oon3m0oo/stack Martin Kroeker 2018-06-20 21:51:38 +0200
  • 05978528c3 Avoid declaring arrays of size 0 when making large stack allocations. Craig Donner 2018-06-20 17:03:18 +0100
  • ef6f0b645e
    Merge pull request #1629 from martin-frbg/issue1628 Martin Kroeker 2018-06-20 16:41:13 +0200
  • 0c5b7b400b
    Add -march=skylake-avx512 to flags if target is skylake x Martin Kroeker 2018-06-20 15:16:19 +0200
  • 4d7719ec34 Further improvements to memory.c. Craig Donner 2018-06-18 17:13:19 +0100
  • 952541e840
    Need to use filter-out to handle NOFORTRAN not set Martin Kroeker 2018-06-20 13:20:30 +0200
  • 9369d3e6e5
    Modify NOFORTRAN tests to always check the value; fix rewriting of NO_FORTRAN Martin Kroeker 2018-06-19 23:28:06 +0200
  • 10b70c904d
    Handle erroneous user settings NOFORTRAN=0 and NO_FORTRAN Martin Kroeker 2018-06-19 20:53:19 +0200
  • 6a5ab083b7
    Handle special case of gfortran+clang+OpenMP Martin Kroeker 2018-06-19 20:47:33 +0200
  • 1f9e4f3193
    Handle special case of gfortran+clang+OpenMP Martin Kroeker 2018-06-19 20:46:36 +0200
  • 5a6a2bed9a
    Merge pull request #1623 from fenrus75/fast-thread Martin Kroeker 2018-06-18 09:02:40 +0200
  • 2d8cc7193a
    Support upcoming Intel Cannon Lake CPUs as Skylake X (#1621) Martin Kroeker 2018-06-17 23:38:14 +0200
  • 2ddc96c9e5 make WMB / MB safer on x86-64 Arjan van de Ven 2018-06-17 18:06:24 +0000
  • 7e39ffe113 On x86-64, make MB/WMB compiler barriers Arjan van de Ven 2018-06-17 17:53:15 +0000
  • 73de17664d Add missing barriers in gemm scheduler Arjan van de Ven 2018-06-17 17:50:43 +0000
  • 6eb4b9ae7c Tune HASWELL SWITCH_RATIO as well Arjan van de Ven 2018-06-17 17:05:04 +0000
  • 5c6f008365 Tune param.h for SkylakeX Arjan van de Ven 2018-06-17 15:47:50 +0000
  • d148ec4ea1 Don't use _Atomic for jobs sometimes... Arjan van de Ven 2018-06-17 15:39:15 +0000
  • 9e162146a9 Only initialize the part of the jobs array that will get used Arjan van de Ven 2018-06-17 15:32:03 +0000
  • 3242317eff
    Fix bad copypasta of code from cpuid_x86 Martin Kroeker 2018-06-17 11:30:25 +0200
  • 2369240a5a
    Support upcoming Cannon Lake as Skylake X Martin Kroeker 2018-06-16 23:13:49 +0200
  • 1dc140ce07
    Support upcoming Cannon Lake as Skylake X Martin Kroeker 2018-06-16 23:13:04 +0200
  • 47bf0dba8f
    Add build-time option for OMP scheduler; document MULTITHREAD_THRESHOLD range (#1620) Martin Kroeker 2018-06-15 11:25:05 +0200
  • b9d1ba2e3a
    Amended description of GEMM_MULTITHREAD_THRESHOLD Martin Kroeker 2018-06-15 09:20:17 +0200
  • 12603b7dbb
    Merge pull request #1618 from oon3m0oo/less_locking Martin Kroeker 2018-06-15 00:10:29 +0200
  • a217049dc6
    Leave DEBUG unset by default Martin Kroeker 2018-06-15 00:09:42 +0200
  • b6b4b3cfc3
    Allow choosing the OpenMP scheduler and add range hint for GEMM_MULTITHREAD_THRESHOLD Martin Kroeker 2018-06-15 00:00:13 +0200
  • b218e1a3c5
    Allow changing the OpenMP scheduler Martin Kroeker 2018-06-14 23:44:33 +0200
  • bf40f806ef Remove the need for most locking in memory.c. Craig Donner 2018-06-14 12:18:04 +0100
  • ed682a4a0c
    Merge pull request #1619 from martin-frbg/issue1580 Martin Kroeker 2018-06-14 17:48:51 +0200
  • fcb77ab129
    Update OSX deployment target to 10.8 Martin Kroeker 2018-06-14 16:57:58 +0200
  • 26e1cfb653
    Merge pull request #1607 from martin-frbg/dynarch Martin Kroeker 2018-06-14 16:52:55 +0200
  • c628c6fa59
    Merge pull request #1612 from oon3m0oo/cpus Martin Kroeker 2018-06-14 16:51:31 +0200
  • 67d81ab49d
    Merge pull request #1609 from martin-frbg/issue1529 Martin Kroeker 2018-06-12 23:00:24 +0200
  • 2f957947a6
    Merge pull request #1613 from xianyi/revert-1600-noyield Martin Kroeker 2018-06-11 17:14:49 +0200
  • de8fff671d
    Revert "Use usleep instead of sched_yield by default" revert-1600-noyield Martin Kroeker 2018-06-11 17:05:27 +0200
  • 6f71c0fce4
    Return a somewhat sane default value for L2 cache size if cpuid retur… (#1611) Martin Kroeker 2018-06-11 13:26:19 +0200
  • c2545b0fd6 Fixed a few more unnecessary calls to num_cpu_avail. Craig Donner 2018-06-11 10:13:09 +0100
  • fc72f68683
    Update setparam-ref.c Martin Kroeker 2018-06-11 09:11:13 +0200
  • 6fa7e8ccd2
    Return a somewhat sane default value for L2 cache size if cpuid returned something unexpected Martin Kroeker 2018-06-11 00:02:05 +0200
  • e65f451409
    include CMakePackageConfigHelpers Martin Kroeker 2018-06-10 15:09:43 +0200
  • 02634b549b
    Add template for OpenBLASConfig.cmake Martin Kroeker 2018-06-10 09:25:46 +0200
  • 0bea6bb9e7
    Create OpenBLASConfig.cmake from cmake as well Martin Kroeker 2018-06-10 09:24:37 +0200
  • 3313e4b946
    Merge pull request #1608 from martin-frbg/issue874 Martin Kroeker 2018-06-09 19:57:33 +0200
  • e9cd11768c
    Enable parallel make on MS Windows by default Martin Kroeker 2018-06-09 17:54:36 +0200
  • 63f7395fb4
    Move some DYNAMIC_ARCH targets to new DYNAMIC_OLDER option Martin Kroeker 2018-06-09 16:31:38 +0200
  • 1cbd8f3ae4
    Move some DYNAMIC_ARCH targets to new DYNAMIC_OLDER option Martin Kroeker 2018-06-09 16:30:46 +0200
  • 6c2d90ba77
    Move some DYNAMIC_ARCH targets to new DYNAMIC_OLDER option Martin Kroeker 2018-06-09 16:29:17 +0200
  • 0297b3211a
    Merge pull request #1605 from oon3m0oo/develop Martin Kroeker 2018-06-09 12:42:34 +0200
  • 66316b9f4c Improve performance of GEMM for small matrices when SMP is defined. Craig Donner 2018-06-07 14:54:42 +0100
  • e09fdbaa06
    Merge 4b0de7690d into 6adc4b7b36 oon3m0oo 2018-06-07 13:57:52 +0000
  • 4b0de7690d Improve performance of GEMM for small matrices when SMP is defined. Craig Donner 2018-06-07 14:54:42 +0100
  • 6adc4b7b36
    Merge pull request #1601 from martin-frbg/zaxpy Martin Kroeker 2018-06-07 14:09:58 +0200
  • 2ade0ef085
    Merge pull request #1600 from martin-frbg/noyield Martin Kroeker 2018-06-07 12:42:00 +0200
  • e8880c1699
    Use a single thread for small input size Martin Kroeker 2018-06-07 10:26:55 +0200
  • ed7c4a043b
    Use usleep instead of sched_yield by default Martin Kroeker 2018-06-07 10:18:26 +0200
  • cf234a0561
    Merge pull request #1589 from fenrus75/skylakex Martin Kroeker 2018-06-06 22:07:09 +0200
  • ae2a33128b
    Merge pull request #1599 from martin-frbg/c_check_avx512 Martin Kroeker 2018-06-06 18:42:42 +0200
  • e4718b1fee
    Better AVX512 test case Martin Kroeker 2018-06-06 16:51:30 +0200
  • 9b87b64262
    Improve AVX512 testcase Martin Kroeker 2018-06-06 16:49:00 +0200
  • 0218b884c1
    Merge pull request #1598 from martin-frbg/issue1593-2 Martin Kroeker 2018-06-06 12:48:26 +0200
  • 83da278093
    Update common.h Martin Kroeker 2018-06-06 09:27:49 +0200
  • 358d4df2bd
    Merge branch 'develop' into issue1593-2 Martin Kroeker 2018-06-06 09:21:41 +0200
  • 06d43760e4
    Restore _Atomic define before stdatomic.h for old gcc Martin Kroeker 2018-06-06 09:18:10 +0200
  • a4af8861ff
    Merge pull request #1597 from martin-frbg/cmake-avx512 Martin Kroeker 2018-06-06 07:22:20 +0200
  • 7fb62aed7e
    Check build system support for AVX512 instructions Martin Kroeker 2018-06-05 23:29:33 +0200
  • 9086d1507b
    Merge a801633194 into ce6317f6c0 Martin Kroeker 2018-06-05 18:29:34 +0000
  • a801633194
    temporarily remove travis_wait Martin Kroeker 2018-06-05 20:29:05 +0200