Didzis Gosko
Didzis Gosko
Updated PR so that for CMake AVX512 must be enabled explicitly. Also added AVX512 subsets for this case as suggested, however I would argue that this will lead to "bad...
Tested inference on a Linux machine with AVX512 all five instruction subsets and AVX512 VNNI.
Hi, this was implemented only for non-Windows systems, but I made an attempt to support Windows platform in the latest commit. I don't have any means to test it myself....
@ggerganov, here it is possible to embed the contents of `cuda-loader.c` into `ggml-cuda.cu` - tested, it works.
I didn't want to step into Windows realm with this PR as it was intended a Linux only feature. Thus I reverted this PR to Linux only solution. Also I...