PR #35231 Enable sdpa micro 4bit kv cache

Enable sdpa micro 4bit kv cache #35231

byungilm wants to merge 8 commits into openvinotoolkit:master from byungilm:enable_sdpa_micro_4bit-kv-cache

Make logic to enable 4bit SDPA KV-cache

8e1530d9

initial sdpa_opt with debug

df2bd636

Reduce cache memory use by using head_size half

d87202c9

Reduce KV-cache memory size halve

8f591438

Resolve accuracy issue if head_size is not multiple of 2*subgroup_size

1c6265f9

Resolve invalid argument index error for SDPA backend execution of sy…

0234e6d2

Minor fix

8df01924

Enable sdpa-micro kernel 4bit KV-cache support

9a820140

github-actions added category: GPU

Reviewers

No reviews

Assignees

No one assigned

Labels

category: GPU

Milestone

No milestone