powerpc: Optimized ZGEMM kernel for POWER10

Browse Source

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.

...

This commit is contained in:

Gordon Fossum

2020-06-24 14:50:12 -05:00

committed by Rajalakshmi Srinivasaraghavan

parent 571eadb880

commit bb2f52844b

4 changed files with 3120 additions and 2 deletions

Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL

Download Patch File Download Diff File Expand all files Collapse all files

kernel/power/zgemm_macros_power10.S

+1138

File diff suppressed because it is too large Load Diff