summaryrefslogtreecommitdiff
path: root/common_power.h
AgeCommit message (Collapse)AuthorFilesLines
2020-11-22Fix syntax mixupMartin Kroeker1-2/+2
2020-11-22Ensure that the same (large) BUFFERSIZE is used for all cpus in DYNAMIC_ARCH ↵Martin Kroeker1-0/+4
builds
2020-10-23Increase BUFFERSIZE furtherMartin Kroeker1-1/+1
2020-10-22Increase BUFFERSIZE for POWER8-10 and use same value for POWER6Martin Kroeker1-2/+2
to fix overflow warning for PWR8 ZGEMM and PWR9 C/ZGEMM and avoid size mismatches in DYNAMIC_ARCH
2020-08-09Add memory barrier to the blas_lock implementation for LinuxMartin Kroeker1-0/+1
as recommended by cparrott73 in #2760
2020-06-11powerpc: Add support for future processorRajalakshmi Srinivasaraghavan1-4/+4
This is the initial patch to support build infrastructure for POWER10 architecture.
2020-04-13Add read barrier definitionMartin Kroeker1-0/+2
2019-12-09Merge pull request #2338 from kavanabhat/aix_modMartin Kroeker1-0/+29
Changes to build on AIX in POWER8 mode
2019-12-06AIX changes for Power8Kavana Bhat1-4/+4
2019-11-03Use the two-operand form of DCBT on all PPC970 regardless of OSMartin Kroeker1-1/+1
There seems to be no advantage to the three-operand form used in the earliest GotoBLAS kernels, and it causes compilation problems on other than the previously special-cased platforms as well
2019-08-20AIX changes for Power8Kavana Bhat1-0/+29
2019-06-28Fix build for PPC970 on FreeBSD pt. 1pkubaj1-1/+1
FreeBSD needs DCBT_ARG=0 as well.
2019-06-25Fix build on FreeBSD/powerpc64.Piotr Kubaj1-3/+3
Signed-off-by: Piotr Kubaj <pkubaj@anongoth.pl>
2019-03-29Merge branch 'develop' into developMartin Kroeker1-3/+13
2019-03-29power9 makefile. dgemm based on power8 kernel with following changes : 32x ↵AbdelRauf1-4/+4
unrolled 16x4 kernel and 8x4 kernel using (lxv stxv butterfly rank1 update). improvement from 17 to 22-23gflops. dtrmm cases were added into dgemm itself
2019-03-25AIX asm syntax changes needed for shared object creationAyappan P1-2/+12
2019-03-07common_power.h: force DCBT_ARG 0 on PPC970 Darwinken-cunningham-webuse1-1/+1
without this, we see ../kernel/power/gemv_n.S:427:Parameter syntax error and many more similar entries that relates to this assembly command dcbt 8, r24, r18 this change makes the DCBT_ARG = 0 and openblas builds through to completion on PowerMac 970 Tests pass
2016-05-16optimized dgemm for 20 threadsWerner Saar1-1/+1
2016-04-21Optimized sgemm and dgemm and tested again.Werner Saar1-0/+5
2016-04-07Updated cgemm- and sgemm-kernel for POWER8 SMPWerner Saar1-1/+1
2016-03-04BUGFIX: increased BUFFER_SIZE for POWER8Werner Saar1-0/+2
2016-03-01added dgemm-, dtrmm-, zgemm- and ztrmm-kernel for power8Werner Saar1-1/+8
2015-08-16add fallback blas_lock implementationGrazvydas Ignotas1-0/+1
to be used on armv5 and new platforms
2015-08-16add fallback rpcc implementationGrazvydas Ignotas1-0/+1
- use on arm, arm64 and any new platform - use faster integer math instead of double - use similar scale as rdtsc so that timeouts work
2015-07-21ppc64le platform support (ELF ABI v2)Matthew Brandyberry1-0/+31
2014-06-27Remove all trailing whitespace except lapack-netlibTimothy Gu1-1/+1
Signed-off-by: Timothy Gu <timothygu99@gmail.com>
2011-01-24Import GotoBLAS2 1.13 BSD version codes.Xianyi Zhang1-0/+795