Skip to content

Pull requests: sgl-project/sgl-kernel-npu

New pull request New
122 Open 571 Closed
122 Open 571 Closed
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
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

feat(kernels-npu): add ELU AscendC operator development sample with tutorial
#788 opened Sep 9, 2026 by cn-sys Loading…
[CI] Migrate A3 runner labels to unified resource pool (752t -> 800t)
#784 opened Sep 8, 2026 by EdwardXuy Contributor Loading…
Pr/chunk kda fwd
#783 opened Sep 8, 2026 by sherdavincl9 • Draft
1
[DeepEP] Fix combine tuning timeout by adding rank synchronization in bench
#781 opened Sep 8, 2026 by 1329009851 Contributor Loading…
4 tasks done
Pr yzy 9 7 2
#780 opened Sep 7, 2026 by YZY00Raiser Contributor Loading…
Lfm
#777 opened Sep 7, 2026 by Hide-on-bushsh Contributor • Draft
ci: re-enable test_mamba_state_update.py in mamba group
#776 opened Sep 4, 2026 by EdwardXuy Contributor Loading…
[feat] Migrate enhanced sparse flash attention from vllm ascend to support dcp for dsa
#775 opened Sep 3, 2026 by heziiop Contributor Loading…
ci: trim redundant A2 single-node deepep test parameters
#772 opened Sep 2, 2026 by EdwardXuy Contributor Loading…
[DeepEP] [feat] Fused moe a5 optimize: Combine two stages into one stage && FP4 weight NZ format supported
#771 opened Sep 2, 2026 by 666syh Contributor Loading…
4 tasks
[DeepEP] [refactor] Merge A5 SiTU epilogue into silu_half with shared…
#770 opened Sep 2, 2026 by WenhaoGe Contributor Loading…
4 tasks done
28
feat(deepep): L2 cross-expert weight prefetch for fused_deep_moe_a5
#765 opened Aug 31, 2026 by X-JD Loading…
[NPU] Fix notify dispatch and mamba causal conv1d dtype
#763 opened Aug 31, 2026 by JinyanYi Loading…
[NPU] Add optimized Kimi-K3 DSpark operators
#762 opened Aug 29, 2026 by McZyWu Contributor • Draft
4 of 6 tasks
Fix fused recurrent state pool indexing
#761 opened Aug 29, 2026 by zihanlin-ai Loading…
1
3
[fix] LoRA sgemmv_shrink_kernel
#759 opened Aug 28, 2026 by Svoloch2940194 Contributor • Draft
[NPU] Add affine-state kernels for Kimi-K3 prefill context parallelism
#758 opened Aug 28, 2026 by qybnb Contributor • Draft
5
Lightning indexer v2 implementation
#757 opened Aug 28, 2026 by TamirBaydasov Loading…
[DeepEP] Add prof for normal kernels
#753 opened Aug 26, 2026 by zuje123 Collaborator Loading…
2 of 4 tasks
[NPU] Disable multibuffer for recompute W/U kernel
#752 opened Aug 26, 2026 by McZyWu Contributor Loading…
1
[feat] add arch35(A5) support for compressor with dual-arch build
#750 opened Aug 26, 2026 by unclezhou486 Contributor Loading…
add chunk gdn kernel and unify ssm state layout for ascend gdn backend
#747 opened Aug 24, 2026 by zhaozx-cn Contributor Loading…
1
[NPU] Add fused Mamba slot copy
#746 opened Aug 24, 2026 by mutian-bd Loading…
perf(kda): add NPU FLA affine-state PCP kernels
#745 opened Aug 24, 2026 by qybnb Contributor Loading…
[NPU] Optimize GDN indexed state pool I/O
#744 opened Aug 24, 2026 by mutian-bd Loading…
ProTip! Find all pull requests that aren't related to any open issues with -linked:issue.