Skip to content

[ET-VK][matmul] Re-implement fp32/fp16 matmul and linear with tiled compute and blocked weight packing#18171

Merged
meta-codesync[bot] merged 5 commits into
gh/SS-JIA/487/basefrom
gh/SS-JIA/487/head
Mar 14, 2026
Merged

[ET-VK][matmul] Re-implement fp32/fp16 matmul and linear with tiled compute and blocked weight packing#18171
meta-codesync[bot] merged 5 commits into
gh/SS-JIA/487/basefrom
gh/SS-JIA/487/head

Commits

Commits on Mar 13, 2026