Skip to content

Pull requests: ROCm/aiter

New pull request New
245 Open 2,068 Closed
245 Open 2,068 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

[quant] refactor hip kernels
#2932 opened Apr 28, 2026 by amd-ruitang3 Contributor Loading…
1 task
1
optimize pa ci:atom
#2930 opened Apr 28, 2026 by fsx950223 Contributor Loading…
1 task
1
Re-enable the native qh128 fp8 kernel (mla_a8w8_qh128_m32x4_n16x2_msk…
#2927 opened Apr 28, 2026 by fangche123 Contributor Loading…
1 task
3
ci: enable prebuilt kernel wheels for fork PRs
#2925 opened Apr 27, 2026 by okakarpa Collaborator Loading…
3 tasks
5
[FLYDSL] Add tuned GDR decode configs
#2924 opened Apr 27, 2026 by xytpai Contributor Loading…
1
Add GLM-4.7 FP8 tuned and untuned FMOE configs
#2923 opened Apr 27, 2026 by omirosh Loading…
2 tasks
5
Deepseek Sparse Attention Triton Kernels for Training
#2922 opened Apr 27, 2026 by wangye805 • Draft
1 task
46
[TRITON] Set RNG seed in Triton tests ci:triton-300x ci:triton-355 triton
#2921 opened Apr 27, 2026 by brunomazzottiamd Contributor Loading…
1 task done
@brunomazzottiamd
2
refactor and unify triton/bench_fav3_sage.py scripts
#2920 opened Apr 27, 2026 by jcaraban Loading…
2 tasks
@juuso-oskari @ksikiric @nsakkine
4
Add paged_attention_ragged_nhd
#2919 opened Apr 27, 2026 by apinge • Draft
1 task
CI: auto-update split test FILE_TIMES
#2918 opened Apr 27, 2026 by github-actions Bot Loading…
2 tasks done
@gyohuangxin
Add a new kernel that supports decode mla with sub_kv=64 and sub_qh=8
#2917 opened Apr 25, 2026 by JohnNikolay84 Contributor Loading…
1 task
@JohnNikolay84
6
Pa gluon swa mtp opt
#2914 opened Apr 25, 2026 by Bernard-Liu Contributor Loading…
1 of 2 tasks
5
[Tracking] Release v0.1.13 validation (DO NOT MERGE) ci:all
#2913 opened Apr 24, 2026 by sunway513 Collaborator Loading…
9 tasks
4
rmsnorm gluon kernel created for gfx1250
#2912 opened Apr 24, 2026 by amd-jrosas Loading…
9
F8 fmha gfx950
#2911 opened Apr 24, 2026 by JohnNikolay84 Contributor Loading…
1 task
@JohnNikolay84
9
improve fused rope kernel
#2910 opened Apr 24, 2026 by amd-weisun Loading…
[MLA] Fix qh128 ASM nullptr write; enable native qh128 fp8 on gfx950
#2907 opened Apr 24, 2026 by inkcherry Contributor Loading…
5
fix edge cases
#2906 opened Apr 24, 2026 by fsx950223 Contributor Loading…
1 task
2
aiter test workflow enhance
#2905 opened Apr 24, 2026 by kiran-thumma Collaborator • Draft
1 task
4
feat(triton/rope): fused QKV split, QK RMSNorm, RoPE, and paged KV cache
#2902 opened Apr 24, 2026 by hellozhuo-amd Loading…
1 task
18
Fix: correct mxfp4 for K not divisible by 256
#2900 opened Apr 24, 2026 by Wanzizhu Loading…
4
Fix CK 2stages MoE (always use BK1 = 16)
#2898 opened Apr 24, 2026 by ex-rzr Loading…
1 task
1
[Bug] Default value of ChunkQ in deepgemm could lead to divided_by_0 error
#2891 opened Apr 24, 2026 by qli88 Loading…
3
Fix 2-stage fused_allreduce_rmsnorm memory ordering
#2890 opened Apr 23, 2026 by hubertlu-tw Contributor Loading…
5 tasks
7
ProTip! Updated in the last three days: updated:>2026-04-24.