Skip to content

Pull requests: vllm-project/vllm-xpu-kernels

Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort

Pull requests list

[ATTN] use b16 policy as fallback path
#574 opened Sep 4, 2026 by YizhouZ Collaborator Loading…
[XPU] Add 2-rank peer-to-peer collectives over Level Zero IPC
#570 opened Sep 3, 2026 by recutita Contributor Loading…
Improve BlockFP8 Xe2 perf by scaling in accumulator
#568 opened Sep 2, 2026 by frost-intel Contributor Loading…
[CONFIGS] add configs for vllm ut
#566 opened Sep 2, 2026 by YizhouZ Collaborator Loading…
update vllm kernel benchmark scripts
#562 opened Sep 1, 2026 by 1pikachu Collaborator Loading…
[XE3P] upstream XE3P kernel
#558 opened Aug 28, 2026 by YizhouZ Collaborator Loading…
[ATTN] head_size padding
#547 opened Aug 25, 2026 by YizhouZ Collaborator Loading…
fuse rmsnorm into mhc_pre & mhc_post_pre
#533 opened Aug 17, 2026 by xiaolong-intel Contributor Loading…
4 tasks
Add INT8 W8A16 (WNA16) linear GEMM kernel
#531 opened Aug 17, 2026 by devan-carlin Loading…
[WIP]bump torch 2.14
#525 opened Aug 14, 2026 by jikunshang Member Loading…
4 tasks
fix(grouped_gemm): initialize Xe2 scheduler counter before launch
#524 opened Aug 13, 2026 by SergiioB Contributor Loading…
[XE2][GDN] Some optimization for gdr prefill kernel
#523 opened Aug 13, 2026 by HeJunyan Contributor Loading…
[XPU] Add block FP8 scale support to grouped GEMM
#490 opened Jul 27, 2026 by kfojcik-intel Loading…
2 of 4 tasks
attn: fail closed on unimplemented softcap and ALIBI
#489 opened Jul 25, 2026 by krisclarkdev Contributor Loading…
1 of 3 tasks
attn: fail closed by default on missing FA2 kernel shapes
#487 opened Jul 25, 2026 by krisclarkdev Contributor Loading…
5 of 7 tasks
attn: enable Split-K for mix-batch decode rows
#485 opened Jul 25, 2026 by krisclarkdev Contributor Loading…
4 tasks done
Repack awq model moe qweight into the fused_moe expects layout
#484 opened Jul 24, 2026 by Yejing-Lai Contributor Loading…
Add Kimi Delta Attention (KDA) and Kimi-Linear Support
#480 opened Jul 23, 2026 by jinyouzhi Loading…
3 of 4 tasks
[XE2][GDN] Optimization for Gated Delta Rule in gdn
#477 opened Jul 23, 2026 by HeJunyan Contributor Loading…
ProTip! no:milestone will show everything without a milestone.