浏览代码

ggml : refactor llamafile_sgemm PPC code (#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 <Shalini.Salomi.Bodapati@ibm.com>
shalinib-ibm 6 月之前
父节点
当前提交
55c509daf5
共有 1 个文件被更改,包括 172 次插入691 次删除
  1. 172 691
      ggml/src/ggml-cpu/llamafile/sgemm.cpp

文件差异内容过多而无法显示
+ 172 - 691
ggml/src/ggml-cpu/llamafile/sgemm.cpp


部分文件因为文件数量过多而无法显示