Commit Graph

4630 Commits

Author SHA1 Message Date
Gordon Fossum
bb2f52844b powerpc: Optimized ZGEMM kernel for POWER10
This patch introduces new optimized version of ZGEMM kernel
using power10 Matrix-Multiply Assist (MMA) feature introduced in
POWER ISA v3.1. This patch makes use of new POWER10 compute instructions
for matrix multiplication operation.

Tested on simulator and there are no new test failures.
Cycles count reduced by 30-50%  compared to POWER9 version depending on
M/N/K sizes.
2020-06-24 14:50:12 -05:00
Rajalakshmi Srinivasaraghavan
571eadb880 powerpc: Optimized SGEMM/DGEMM/CGEMM for POWER10
This patch introduces new optimized version of SGEMM, CGEMM and DGEMM
using power10 Matrix-Multiply Assist (MMA) feature introduced in
POWER ISA v3.1. This patch makes use of new POWER10 compute instructions
for matrix multiplication operation.

Tested on simulator and there are no new test failures.
Cycles count reduced by 30-50%  compared to POWER9 version depending on
M/N/K sizes.
MMA GCC patch for reference:
https://gcc.gnu.org/git/?p=gcc.git;a=commit;h=8ee2640bfdc62f835ec9740278f948034bc7d9f1
2020-06-24 14:48:15 -05:00
Martin Kroeker
93592d1260 Merge pull request #2675 from wjc404/develop
AVX512 DGEMM TCOPY_16 Function
2020-06-23 09:29:02 +02:00
Martin Kroeker
6eaeb01263 Merge pull request #2658 from RajalakshmiSR/p10
powerpc: Add support for future processor
2020-06-23 00:02:37 +02:00
wjc404
086d87a302 AVX512 dgemm tcopy_16 function 2020-06-20 00:07:43 +08:00
Martin Kroeker
af501eb753 Merge pull request #2669 from mhillenibm/zarch_fix_gcc_detection
Zarch fix gcc detection
2020-06-17 17:55:25 +02:00
Martin Kroeker
0eb6c4dded Merge pull request #2672 from mhillenibm/test_num_threads
cpp_thread_test: Change adjustment of concurrency on systems with <52 hw threads
2020-06-17 17:54:31 +02:00
Marius Hillenbrand
de838c38ef cpp_thread_test/dgemv: fail early if concurrency is zero
The two test cases dgemv_tester and dgemm_tester accept the degree of
concurrency as command line argument (amongst others). Fail early if
value 0 has been specified, instead of later with less-clear symptoms.

Signed-off-by: Marius Hillenbrand <mhillen@linux.ibm.com>
2020-06-17 16:15:44 +02:00
Marius Hillenbrand
478898b37a cpp_thread_test/dgemv: cap concurrency to number of hw threads on small systems
... instead of (number of hw threads - 4) to avoid invalid numbers on
smaller systems. Currently, systems with 4 or fewer CPUs (e.g., small CI
VMs) would fail the test. Fixes one of the issues discussed in #2668

Signed-off-by: Marius Hillenbrand <mhillen@linux.ibm.com>
2020-06-17 16:08:48 +02:00
Marius Hillenbrand
2389291766 Makefile.system: remove duplicate variable GCCVERSIONGT5
... to bring unified gcc version detection with common variables to the
one remaining spot in Makefile.system.

Signed-off-by: Marius Hillenbrand <mhillen@linux.ibm.com>
2020-06-16 15:06:03 +02:00
Marius Hillenbrand
a2d13ea611 Fix gcc version detection for zarch
Employ common variables for gcc version detection and fix the broken
check for gcc >= 5.2.
Fixes #2668

Signed-off-by: Marius Hillenbrand <mhillen@linux.ibm.com>
2020-06-16 15:06:03 +02:00
Martin Kroeker
1bd3cd66c2 Increment version to 0.3.10.dev 2020-06-14 22:05:19 +02:00
Martin Kroeker
1c53e1366d Increment version to 0.3.10.dev 2020-06-14 22:04:37 +02:00
Martin Kroeker
95dbeff66d Merge branch 'release-0.3.0' into develop 2020-06-14 22:02:45 +02:00
Martin Kroeker
3b673a24b7 Increment version to 0.3.10.dev 2020-06-14 21:57:52 +02:00
Martin Kroeker
1eb1979050 Increment version to 0.3.10.dev 2020-06-14 21:57:15 +02:00
Martin Kroeker
efc53b6e7e Merge pull request #2665 from martin-frbg/flang-fixes-2a
Fix spelling of flang option -Mrecursive, add -Kieee and workaround for AOCC optimizer bug
2020-06-14 21:56:08 +02:00
Martin Kroeker
72888497e2 Update with 0.3.10 changes 2020-06-14 21:55:31 +02:00
Martin Kroeker
7e3e006af6 Merge pull request #2666 from martin-frbg/blastest
Update BLAS tests to what netlib 3.9.0 uses
2020-06-14 18:28:37 +02:00
Martin Kroeker
d906d14402 Merge pull request #2664 from ACSimon33/exported_symbols
Add missing exported symbols.
2020-06-14 18:27:03 +02:00
Martin Kroeker
3785c0e82b Merge pull request #2663 from martin-frbg/issue2654
Respect predefined defaults for AR, AS, LD and RANLIB
2020-06-14 18:26:43 +02:00
Martin Kroeker
f2d8879af6 Merge pull request #2661 from martin-frbg/issue2660
Report selected DYNAMIC_ARCH kernel rather than one of its aliases in gotoblas_corename
2020-06-14 18:25:37 +02:00
Martin Kroeker
6876221cf3 Remove optimization level limit for flang again and add -fno-unroll-loops for AOCC flang 2.x instead 2020-06-14 17:40:24 +02:00
Martin Kroeker
79cdcde717 Re-enable higher optimization levels for flang while disabling loop unrolling for AOCC flang 2020-06-14 17:18:16 +02:00
Martin Kroeker
18a11137f1 Update BLAS tests to correspond to Reference-LAPACK 3.9.0
replaces calculation of machine precision with call to epsilon intrinsic and removes the requirement for previous output files to be removed before rerunning tests
2020-06-14 10:26:25 +02:00
Martin Kroeker
1dd712131e Fix spelling of flang option -Mrecursive and add -Kieee 2020-06-14 00:09:31 +02:00
Martin Kroeker
0ed2adf0b2 Fix spelling of flang option -Mrecursive and add -Kieee 2020-06-14 00:01:20 +02:00
Martin Kroeker
abf670757b Respect predefined defaults for AR, AS, LD and RANLIB 2020-06-13 23:21:13 +02:00
Simon Märtens
41fc6f3cd2 Added missing exported symbols. 2020-06-13 22:37:39 +02:00
Martin Kroeker
007d9f97d7 Make gotoblas_corename report the name of the selected TARGET rather than its aliases 2020-06-13 19:25:28 +02:00
Martin Kroeker
63d26090f5 Merge pull request #64 from xianyi/develop
rebase
2020-06-13 19:14:47 +02:00
Rajalakshmi Srinivasaraghavan
9fe930f205 powerpc: Add support for future processor
This is the initial patch to support build infrastructure
for POWER10 architecture.
2020-06-11 15:47:20 -05:00
Martin Kroeker
3a1b58d54a Merge pull request #2653 from craft-zhang/cortex-a53
fix INIT8x4 of SGEMM on Arm Cortex-A53
2020-06-10 12:19:33 +02:00
Martin Kroeker
f7659be4a0 Merge pull request #2652 from martin-frbg/flang-fixes
Fixes for compilation with flang binary release 20190329
2020-06-09 20:31:06 +02:00
ZhangDanfeng
bc6fd20a40 fix INIT8x4
Signed-off-by: ZhangDanfeng <467688405@qq.com>
2020-06-10 01:01:16 +08:00
Martin Kroeker
3ce469a34f Limit optimization level to O1 for flang and add -frecursive 2020-06-09 16:11:13 +02:00
Martin Kroeker
ba2c5b404d When building with flang, use it also for the final link step to get dependencies right 2020-06-09 16:09:34 +02:00
Martin Kroeker
f07a80354b Apply previously AOCC-specific workaround to all versions of flang 2020-06-09 16:07:03 +02:00
Martin Kroeker
fdd1b50263 Merge pull request #63 from xianyi/develop
rebase
2020-06-09 15:54:30 +02:00
Martin Kroeker
430e8b45fe Merge pull request #2648 from martin-frbg/lapack411
Break out of potentially infinite rescaling loop in LAPACK xLARGV/xLARTG/xLARTGP
2020-06-07 19:45:52 +02:00
Martin Kroeker
88fe85f4e0 Merge pull request #2647 from martin-frbg/aocc-flang
Small fixes for flang in general and the AMD AOCC version of it in particular
2020-06-07 19:45:11 +02:00
Martin Kroeker
89091e6b64 Merge pull request #2645 from martin-frbg/misc_fixes
Miscellaneous fixes
2020-06-07 19:44:50 +02:00
Martin Kroeker
522aaf53bf Break out of potentially infinite rescaling loop in LAPACK xLARGV/xLARTG/xLARTGP
Reference-LAPACK issue 411
2020-06-07 14:30:20 +02:00
Martin Kroeker
c3574ffe53 Merge pull request #2646 from wjc404/develop
Optimize AVX512 parallel DGEMM performance
2020-06-07 13:18:22 +02:00
Martin Kroeker
4e28dc6353 Use only -O1 with AMD AOCC version of flang
to prevent miscompilation of LAPACK codes and tests on Ryzen
2020-06-07 00:05:02 +02:00
Martin Kroeker
13c28889a2 Update "cosmetic fixes for non-C99 compilers" 2020-06-06 15:22:27 +02:00
wjc404
0e3ac4a06b Add files via upload 2020-06-06 14:56:57 +08:00
Martin Kroeker
28915eed72 Cosmetic fixes for non-C99 compilers 2020-06-05 10:05:34 +02:00
Martin Kroeker
7f60fb6b91 Delete spurious copy of common_param.h 2020-06-05 10:04:16 +02:00
Martin Kroeker
0464e662ad make blas_quickdivide unsigned and guard against miscompilation 2020-06-05 10:03:36 +02:00