Martin Kroeker
3eec7d382c
ARMV7 does not support DMB ISHLD, use DMB ISH
2020-04-13 15:56:31 +02:00
Martin Kroeker
8692456226
Add read barrier definition
2020-04-13 12:10:37 +02:00
Martin Kroeker
a33d177430
Increase default BUFFER_SIZE on ARM, ZARCH and newer x86_64, add GEMM_R for POWER8/9
...
As shown in #2538 , default buffersizes on some platforms were smaller than required in memory.c
and the requirement could never be fulfilled for a calculated GEMM_R on PPC given the fomula used
2020-04-12 19:44:48 +02:00
Ashwin Sekhar T K
0d5c8e5386
arm: Determine the abi from compiler if not specified on command line
...
If ARM abi is not explicitly mentioned on the command line, then set the
arm abi to softfp or hard according to the compiler environment.
This assumes that compiler sets the defines __ARM_PCS and __ARM_PCS_VFP
accordingly.
2017-06-30 18:20:59 +05:30
Zhang Xianyi
503dcbfde6
Merge branch 'develop' into arm_soft_fp_abi
2017-03-06 13:53:56 +08:00
Ivan Shapovalov
6c0862a94f
common_arm.h, common_mips.h: get rid of .func directives
...
.func/.endfunc are gcc/gas-specific directives for generating stabs
debug information (and nothing more). This is near-useless now because
DWARF is commonly used, and not implemented in Clang. Hence building
OpenBLAS with Clang fails, and there is no sane way to detect GCC vs.
anything else with preprocessor definitions.
Hence, just remove these directives.
2016-09-09 03:37:11 +03:00
Zhang Xianyi
857899526f
ARM soft fp abi branch.
2015-09-26 14:10:18 +00:00
Grazvydas Ignotas
d38a1ddc7a
use real armv5 support
...
there is no more requirement for ARMv6 instructions,
and VFP on ARMv5 is uncommon
2015-08-16 18:59:18 +02:00
Grazvydas Ignotas
6b92204a7c
add fallback blas_lock implementation
...
to be used on armv5 and new platforms
2015-08-16 18:59:17 +02:00
Grazvydas Ignotas
f2ac1a5cee
set ARMV7 for Cortex-A9 and Cortex-A15
...
otherwise some macros like YIELDING are not defined correctly
2015-08-16 18:59:17 +02:00
Grazvydas Ignotas
e12cf1123e
add fallback rpcc implementation
...
- use on arm, arm64 and any new platform
- use faster integer math instead of double
- use similar scale as rdtsc so that timeouts work
2015-08-16 18:59:16 +02:00
Grazvydas Ignotas
c2323dd4d2
really fix ARM locking
...
- was writing 0 to lock variable, so was ineffective
- only exit loop if both lock was 0 and strex was successful
2015-08-16 15:18:42 +02:00
Werner Saar
eea2e30b74
bugfix for arm locking
2015-05-23 11:40:40 +02:00
Werner Saar
ea4df0aad3
Ref #574 : Bugfix for armv6 memory barrier
2015-05-19 10:43:12 +02:00
Werner Saar
3814bf60d3
added optimized dsymv kernels for haswell
2015-04-22 10:42:50 +02:00
Zhang Xianyi
2fb02626da
Update organization info.
2014-11-25 15:28:58 +08:00
Timothy Gu
6c2ead30f0
Remove all trailing whitespace except lapack-netlib
...
Signed-off-by: Timothy Gu <timothygu99@gmail.com >
2014-06-27 12:05:18 -07:00
wernsaar
5400a9f4e4
redefined functions for TIMING and YIELDING for ARMV7 processor
2013-11-03 10:34:04 +01:00
wernsaar
02bc36ac79
added sgemm_ncopy routine and made some improvements on cgemm_kernel for ARMV7
2013-11-01 18:22:27 +01:00
wernsaar
d13788d1b4
common files modified for ARM
2013-09-28 19:10:32 +02:00
