5b95534afc
Make TARGET=GENERIC compatible with DYNAMIC_ARCH=1
...
for issue #2048
2019-03-09 11:21:16 +01:00
b7f59da42d
Fix crash in sgemm SSE/nano kernel on x86_64
...
Fix bug #2047 .
Signed-off-by: Celelibi <celelibi@gmail.com>
2019-03-07 16:55:13 +01:00
783ba8058f
HiSilicon tsv110 CPUs optimization branch
...
add HiSilicon tsv110 CPUs optimization branch
2019-03-04 16:30:50 +08:00
6eee1beac5
move fix to right place
2019-02-24 20:41:02 +02:00
e12cdf58ef
Merge pull request #2024 from martin-frbg/gcc9fixes4
...
Fix inline assembly constraints in Bulldozer TRSM kernels
2019-02-17 11:49:15 +01:00
1860c9456d
Merge pull request #2023 from martin-frbg/gcc9fixes3
...
Fix inline assembly constraints in various x86_64 GEMVN kernels
2019-02-17 11:48:57 +01:00
f9bb76d29a
Fix inline assembly constraints in Bulldozer TRSM kernels
...
rework indices to allow marking i,as and bs as both input and output (marked operand n1 as well for simplicity). For #2009
2019-02-16 20:06:48 +01:00
efb9038f72
Fix inline assembly constraints
2019-02-16 18:46:17 +01:00
e976557d29
Fix inline assembly constraints
...
rework indices to allow marking argument lda as input and output.
2019-02-16 18:36:39 +01:00
9d8be15789
Fix inline assembly constraints
...
rework indices to allow marking argument lda4 as input and output. For #2009
2019-02-16 18:24:11 +01:00
d752799a0f
Merge pull request #2021 from martin-frbg/gcc9fixes2
...
Fix wrong constraints in inline assembly of Haswell DTRSM kernel
2019-02-16 18:05:40 +01:00
c26c0b77a7
Fix wrong constraints in inline assembly
...
for #2009
2019-02-15 15:08:16 +01:00
1c6da2d03c
Merge pull request #2019 from martin-frbg/gcc9fixes
...
Fix unannounced modification of input operand 8 (lda4) in Haswell GEMVN microkernel
2019-02-15 15:02:54 +01:00
4255a58cd2
Rename operands to put lda on the input/output constraint list
2019-02-15 10:10:04 +01:00
46e415b140
Save and restore input argument 8 (lda4)
...
Fixes miscompilation with gcc9 -ftree-vectorize (related to issue #2009 )
2019-02-14 22:43:18 +01:00
69a97ca7b9
dgemv_kernel_4x4(Haswell): add missing clobbers for xmm0,xmm1,xmm2,xmm3
...
This fixes a crash in dblat2 when OpenBLAS is compiled using
-march=znver1 -ftree-vectorize -O2
See also:
https://github.com/easybuilders/easybuild-easyconfigs/issues/7180
2019-02-14 16:27:58 +00:00
056917d616
Merge pull request #2013 from martin-frbg/issue2011
...
Fix invalid memory access in PPC gemm_beta
2019-02-14 09:29:34 +01:00
718efcec6f
Fix out-of-bounds memory access in gemm_beta
...
Fixes #2011 (as suggested by davemq), assuming typo by K.Goto
2019-02-13 22:08:37 +01:00
f9d67bb5e8
Fix out-of-bounds memory access in gemm_beta
...
Fixes #2011 (as suggested by davemq) presuming typo by K.Goto
2019-02-13 22:06:41 +01:00
76bb74fcd4
Merge pull request #2012 from maamountki/z14
...
[ZARCH] Many improvements
2019-02-13 20:15:56 +01:00
0a54c98b9d
[ZARCH] Modify constraints
2019-02-13 21:06:25 +02:00
bec54ae366
[ZARCH] Fix caxpy
2019-02-13 12:54:35 +02:00
ab1630f9fa
Fix declaration of arguments in inline assembly
...
Argument 0 is modified so should be input and output
2019-02-12 16:14:02 +01:00
b824fa70eb
Fix declaration of assembly arguments in SSYMV and DSYMV microkernels
...
Arguments 0 and 1 are both input and output
2019-02-12 16:00:18 +01:00
91481a3e4e
Fix declaration of input arguments in inline assembly
...
Argument 0 is modified as it doubles as a counter
2019-02-12 15:51:43 +01:00
dc6ac9eab0
Fix declaration of input arguments in the x86_64 s/dGEMV_T and s/dGEMV_N kernels
...
Arguments 0 and 1 need to be tagged as both input and output
2019-02-12 15:33:48 +01:00
f583674109
[ZARCH] Fix cgemv_t_4
2019-02-12 13:12:28 +02:00
77fe70019f
[ZARCH] Fix constraints and source code formatting
2019-02-11 16:01:13 +02:00
7039770165
[ZARCH] Undo the last commit
2019-02-06 20:11:44 +02:00
11a43e8116
[ZARCH] Set alignment hint for vl/vst
2019-02-05 19:17:08 +02:00
61526480f9
[ZARCH] Fix copy constraint
2019-02-05 07:51:19 +02:00
81daf6bc38
[ZARCH] Format source code, Fix constraints
2019-02-05 07:30:38 +02:00
729e925174
Merge pull request #1996 from quickwritereader/develop
...
NBMAX=4096 for gemvn, added sgemvn 8x8 for future
2019-02-04 16:52:04 +01:00
498ac98581
Note for unused kernels
2019-02-04 15:41:56 +00:00
cd9ea45463
NBMAX=4096 for gemvn, added sgemvn 8x8 for future
2019-02-04 06:57:11 +00:00
f9c5023e04
Merge pull request #1994 from quickwritereader/develop
...
sgemv cgemv pairs
2019-02-01 21:04:47 +01:00
4abc375a91
sgemv cgemv pairs
2019-02-01 13:45:00 +00:00
874df65491
Fix incorrect sgemv results for IBM z14
...
part of PR #1993 that was inadvertently misplaced into the toplevel directory
2019-02-01 12:58:59 +01:00
877023e1e1
Fix precision of zarch DSDOT
...
from patch provided by aarnez in #991
2019-01-31 21:22:26 +01:00
265142edd5
Fix typo in the zarch min/max kernels
...
from patch provided by aarnez in #991
2019-01-31 21:21:40 +01:00
885a3c4350
USE_TRMM on Z14
...
from patch provided by aarnez in #991
2019-01-31 21:18:09 +01:00
82124729af
Merge branch 'develop' into z14
2019-01-31 19:36:41 +02:00
29416cb5a3
[ZARCH] Add Z13 version for max/min functions
2019-01-31 19:11:11 +02:00
48b9b94f7f
[ZARCH] Improve loading performance for camax/icamax
2019-01-31 18:52:11 +02:00
86a824c97f
Fix wrong comparison that made IMIN identical to IMAX
...
as reported by aarnez in #1990
2019-01-31 15:27:21 +01:00
808410c2c7
Fix wrong comparison that made IMIN identical to IMAX
...
as suggested in #1990
2019-01-31 15:25:15 +01:00
fcd814a8d2
[ZARCH] Fix bug in max/min functions
2019-01-29 17:59:38 +02:00
dc4d3bccd5
[ZARCH] Fix icamax/icamin
2019-01-29 03:47:49 +02:00
c7143c1019
[ZARCH] Fix iamax/imax single precision
2019-01-28 17:52:23 +02:00
04873bb174
[ZARCH] Undo the last commit
2019-01-28 17:32:24 +02:00
c8ef9fb220
[ZARCH] Fix bug in iamax/iamin/imax/imin
2019-01-28 17:16:18 +02:00
b111829226
[ZARCH] Update max/min functions
2019-01-21 15:56:04 +02:00
32b0f1168e
Fix declaration of input arguments in the Sandybridge GER microkernels ( #1967 )
...
* Tag arguments 0 and 1 as both input and output
2019-01-18 08:11:39 +01:00
b495e54310
Fix declaration of input arguments in the x86_64 SCAL microkernels ( #1966 )
...
* Tag arguments 0 and 1 as both input and output (see #1964 )
2019-01-18 08:11:07 +01:00
d5e6940253
Fix declaration of input arguments in the x86_64 microkernels for DOT and AXPY ( #1965 )
...
* Tag operands 0 and 1 as both input and output
For #1964 (basically a continuation of coding problems first seen in #1292 )
2019-01-17 23:20:32 +01:00
43a4572038
crot fix
2019-01-17 14:45:31 +00:00
a034e65512
Merge branch 'develop' into develop
2019-01-16 19:25:13 +04:00
8c3386be87
Added missing Blas1 single fp {saxpy, caxpy, cdot, crot(refactored version of srot),isamax ,isamin, icamax, icamin},
...
Fixed idamin,icamin choosing the first occurance index of equal minimals
2019-01-16 15:16:21 +00:00
b815a04c87
[ZARCH] fix a bug in max/min functions
2019-01-15 21:04:22 +02:00
1a7925b3a3
[ZARCH] Update dgemv_n_4.c
2019-01-11 17:43:11 +02:00
406f835f00
[ZARCH] update cgemv_n_4.c
2019-01-11 17:39:17 +02:00
621dedb37b
[ZARCH] Update cgemv_t_4.c
2019-01-11 17:37:11 +02:00
b731e8246f
Update sgemv_t_4.c
2019-01-11 17:14:04 +02:00
ecc31b743f
Update dgemv_t_4.c
2019-01-11 17:13:02 +02:00
5d89d6b143
[ZARCH] fix sgemv_n_4.c
2019-01-11 17:08:24 +02:00
67432b23c2
[ZARCH] fix cgemv_n_4.c
2019-01-11 16:44:46 +02:00
be66f5d5c2
[ZARCH] fix data prefetch type in sdot
2019-01-09 16:50:07 +02:00
c2ffef8156
[ZARCH] fix data prefetch type in ddot
2019-01-09 16:49:44 +02:00
e7455f500c
[ZARCH] fix dsdot.c
2019-01-09 16:33:54 +02:00
3eafcfa650
[ZARCH] fix cgemv_n_4.c
2019-01-09 07:43:45 +02:00
94cd946b96
[ZARCH] fix cgemv_n_4.c
2019-01-04 17:45:56 +02:00
1aa840a0a2
[ZARCH] fix sgemv_t_4.c
2019-01-04 01:38:18 +02:00
795285c587
Fix thinko in skylake beta handling
...
casting ints is cheaper but it has a rounding, not memory casing effect, resulting in
invalid outcome
2018-12-24 18:49:50 +00:00
d321448a63
dgemm: use dgemm_ncopy_8_skylakex.c also for Haswell
...
The dgemm_ncopy_8_skylakex.c code is not avx512 specific and gives
a nice performance boost for medium sized matrices
2018-12-16 23:09:22 +00:00
c43331ad0a
dgemm: Use the skylakex beta function also for haswell
...
it's more efficient for certain tall/skinny matrices
2018-12-16 23:09:17 +00:00
c4e23dd016
Update Makefile
2018-12-16 18:14:40 +01:00
cfc4acc221
typo
2018-12-16 16:19:51 +01:00
545c2b1bbb
Add -mavx2 on Haswell only if the compiler supports it
2018-12-16 13:09:19 +01:00
69d206440a
Make the skylakex/haswell sgemm code compile and run even with compilers without avx2 support
2018-12-16 00:19:41 +00:00
3843e3e017
use -maxv2 on haswell
2018-12-15 23:30:31 +01:00
fbcb14a74b
should be core-avx2
2018-12-15 20:18:59 +01:00
2a3190dc76
fix elseifeq and use older option core2-avx for compatibility
2018-12-15 20:17:44 +01:00
1ebe5c0f49
Add -march=haswell to HASWELL part of DYNAMIC_ARCH build
2018-12-15 19:35:35 +01:00
0586899a10
Use sgemm_ncopy_4_skylakex.c also for Haswell
...
sgemm_ncopy_4_skylakex.c uses SSE transpose operations where the
real perf win happens; this also works great for Haswell.
This gives double digit percentage gains on small and skinny matrices
2018-12-15 13:49:19 +00:00
00dc09ad19
Use the skylake sgemm beta code also for haswell
...
with a few small changes it's possible to use the skylake sgemm code
also for haswell, this gives a modest gain (10% range) for smallish
matrixes but does wonders for very skinny matrixes
2018-12-15 13:49:13 +00:00
cdc668d82b
Add a "sgemm direct" mode for small matrixes
...
OpenBLAS has a fancy algorithm for copying the input data while laying
it out in a more CPU friendly memory layout.
This is great for large matrixes; the cost of the copy is easily
ammortized by the gains from the better memory layout.
But for small matrixes (on CPUs that can do efficient unaligned loads) this
copy can be a net loss.
This patch adds (for SKYLAKEX initially) a "sgemm direct" mode, that bypasses
the whole copy machinary for ALPHA=1/BETA=0/... standard arguments,
for small matrixes only.
What is small? For the non-threaded case this has been measured to be
in the M*N*K = 28 * 512 * 512 range, while in the threaded case it's
less, around M*N*K = 1 * 512 * 512
2018-12-13 13:47:31 +00:00
87718807f0
Merge pull request #1910 from martin-frbg/issue1909
...
Fix for DYNAMIC_ARCH builds made on a AVX512-capable host
2018-12-12 14:56:25 +01:00
51aec8e96b
make sure the added march=skylake-avx512 does not cause problems on Windows
2018-12-11 22:47:32 +01:00
06f7d78d70
Add -march=skylake-avx512 to SkylakeX part of DYNAMIC_ARCH builds
2018-12-11 21:10:38 +01:00
7639f2e1f0
Rewrite the conditional for OSX to fix cmake parsing on others
...
The Makefile variable parser in utils.cmake currently does not handle conditionals. Having the definitions for non-OSX last will at least make cmake builds work again on non-OSX platforms.
2018-12-06 14:04:27 +01:00
2fc712469d
Avoid creating spurious non-suffixed c/zgemm_kernels
...
Plain cgemm_kernel and zgemm_kernel are not used anywhere, only cgemm_kernel_b etc.
Needlessly building them (without any define like NN, CN, etc.) just happened to work on most platforms, but not on arm64. See #1870
2018-12-06 13:56:06 +01:00
6ba30e270d
Fix typo that broke CNRM2 on ARMV8 since 0.3.0
...
must have happened in my #1449
2018-12-06 13:42:25 +01:00
701ea88347
Use p2align instead of align for OSX compatibility
...
fixes #1902
2018-12-03 13:06:43 +01:00
6c7b691083
Really revert xDOT changes from 1832
...
neglected to rebase #1892 on merging
2018-11-30 21:32:01 +01:00
5f4c550c27
Merge pull request #1892 from martin-frbg/mipsdot
...
revert MIPS64 xDOT kernel changes from #1832
2018-11-30 21:28:21 +01:00
95a5542e3c
Revert DOT kernel changes from #1834
...
as the failures seen on Loongson3A appear to be limited to DSDOT/SDSDOT (i.e. my hackish "fix" from #1684 )
2018-11-30 11:16:24 +01:00
7a2e1bc804
Use generic kernel for DSDOT/SDSDOT
...
as discussed in #1834
2018-11-30 10:57:09 +01:00
35653e38b3
Merge pull request #1834 from fengrl/develop
...
register push/pop command change
2018-11-30 10:48:46 +01:00
19c4bdd8b3
Add return value so that freebsd system clang does not err out
2018-11-25 21:35:01 +01:00
310ea55f29
Simplifying ARMv8 build parameters
...
ARMv8 builds were a bit mixed up, with ThunderX2 code in ARMv8 mode
(which is not right because TX2 is ARMv8.1) as well as requiring a few
redundancies in the defines, making it harder to maintain and understand
what core has what. A few other minor issues were also fixed.
Tests were made on the following cores: A53, A57, A72, Falkor, ThunderX,
ThunderX2, and XGene.
Tests were: OpenBLAS/test, OpenBLAS/benchmark, BLAS-Tester.
A summary:
* Removed TX2 code from ARMv8 build, to make sure it is compatible with
all ARMv8 cores, not just v8.1. Also, the TX2 code has actually
harmed performance on big cores.
* Commoned up ARMv8 architectures' defines in params.h, to make sure
that all will benefit from ARMv8 settings, in addition to their own.
* Adding a few more cores, using ARMv8's include strategy, to benefit
from compiler optimisations using mtune. Also updated cache
information from the manuals, making sure we set good conservative
values by default. Removed Vulcan, as it's an alias to TX2.
* Auto-detecting most of those cores, but also updating the forced
compilation in getarch.c, to make sure the parameters are the same
whether compiled natively or forced arch.
Benefits:
* ARMv8 build is now guaranteed to work on all ARMv8 cores
* Improved performance for ARMv8 builds on some cores (A72, Falkor,
ThunderX1 and 2: up to 11%) over current develop
* Improved performance for *all* cores comparing to develop branch
before TX2's patch (9% ~ 36%)
* ThunderX1 builds are 14% faster than ARMv8 on TX1, 9% faster than
current develop's branch and 8% faster than deveop before tx2 patches
Issues:
* Regression from current develop branch for A53 (-12%) and A57 (-3%)
with ARMv8 builds, but still faster than before TX2's commit (+15%
and +24% respectively). This can be improved with a simplification of
TX2's code, to be done in future patches. At least the code is
guaranteed to be ARMv8.0 now.
Comments:
* CortexA57 builds are unchanged on A57 hardware from develop's branch,
which makes sense, as it's untouched.
* CortexA72 builds improve over A57 on A72 hardware, even if they're
using the same includes due to new compiler tunning in the makefile.
2018-11-19 16:41:49 +00:00
Martin Kroeker
Celelibi
maomao194313
Andrew
Bart Oldeman
maamountki
Ubuntu
Abdelrauf
Arjan van de Ven
Renato Golin