xla
Add heuristic default block sizes for different cases in ragged attention kernel
#8922
Merged

Add heuristic default block sizes for different cases in ragged attention kernel #8922

yaochengji merged 8 commits into master from chengji/ragged-attn
yaochengji
yaochengji Add heuristic default block sizes for different cases in ragged atten…
64ac8b68
yaochengji yaochengji requested a review from vanbasten23 vanbasten23 256 days ago
bythew3i
bythew3i commented on 2025-04-02
vanbasten23
vanbasten23 commented on 2025-04-02
vanbasten23
vanbasten23 commented on 2025-04-02
yaochengji fix
437fa7d9
yaochengji fix
e4a23760
yaochengji fix
3e54ecbe
yaochengji fix
faa2d2a1
yaochengji fix block size
56147e0f
vanbasten23
vanbasten23 approved these changes on 2025-04-02
yaochengji set v4 of vmem_limit_bytes to 16MB
8286d566
yaochengji yaochengji force pushed from 64d9bad4 to 8286d566 254 days ago
yaochengji fix
119a50b6
yaochengji yaochengji merged 6c3f2313 into master 254 days ago

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
Milestone