]> git.djapps.eu Git - pkg/ggml/sources/ggml/commit
ggml-cpu: add always_inline to tinyBLAS_PPC accumulator saves (llama/20791)
authorshalinib-ibm <redacted>
Fri, 20 Mar 2026 23:11:45 +0000 (04:41 +0530)
committerGeorgi Gerganov <redacted>
Sat, 28 Mar 2026 11:39:09 +0000 (13:39 +0200)
commit9b6c70190e6b9df0423604dccda24c4162db9aef
treef80882b2fd2ddd3161db30d78d19aab398e22f44
parentbb9309adc4cbef6733cd004469ccd8966b22bfd1
ggml-cpu: add always_inline to tinyBLAS_PPC accumulator saves (llama/20791)

Explicitly mark save_acc and add_save_Acc with always_inline
in tinyBLAS_PPC. This ensures the compiler keeps MMA accumulator
disassembly within kernel's register context, preventing un-necessary
stask spills.

Signed-off-by: Shalini Salomi Bodapati <redacted>
src/ggml-cpu/llamafile/sgemm.cpp