]> git.djapps.eu Git - pkg/ggml/sources/whisper.cpp/commit
ggml : refactor llamafile_sgemm PPC code (llama/14673)
authorshalinib-ibm <redacted>
Mon, 14 Jul 2025 13:16:42 +0000 (18:46 +0530)
committerGeorgi Gerganov <redacted>
Sat, 19 Jul 2025 21:23:50 +0000 (00:23 +0300)
commitded2e3cf6d29838b83cf74f3ae4d407f895cf12d
tree90722bd2c9920c263a17dbd3e752e67e66e408fa
parentebb0e9d0ed7eccd5b320d56d60ef50cf3c87b105
ggml : refactor llamafile_sgemm PPC code (llama/14673)

Remove un-necessary templates from class definition and packing functions
Reduce deeply nested conditionals, if-else switching in mnapck function
Replace repetitive code with inline functions in Packing functions

2 ~ 7% improvement in Q8 Model
15 ~ 50% improvement in Q4 Model

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