Skip to content

Pull requests: ROCm/aiter

New pull request New
532 Open 3,925 Closed
532 Open 3,925 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

[TRITON][GLUON][GFX950] Sparse paged attention and Sparse MLA
#4919 opened Aug 21, 2026 by cagrikymk Contributor • Draft
2 tasks
4
[Triton/Gluon] Retune unified attention configs for gfx950 ci:triton-300x ci:triton-355
#4918 opened Aug 21, 2026 by nidal567 Contributor Loading…
1 task done
@nidal567
1
[Triton/Gluon] Move gluon gemm_a8w8_blockscale kernel into _gluon_kernels
#4917 opened Aug 21, 2026 by vgokhale Contributor • Draft
9
[Bugfix] Fix ASM split-K semaphore deadlock under CUDA graph capture
#4916 opened Aug 21, 2026 by JohnQinAMD Contributor Loading…
4 tasks done
5
[Config] [Kimi K3 fix] Drop gfx942/gfx950 opus rows from the BF16 tuned-GEMM tables
#4915 opened Aug 21, 2026 by hyukjlee Loading…
1
[CI] [Build] chore(flydsl): bump flydsl dependency to 0.3.2.dev838
#4912 opened Aug 21, 2026 by coderfeli Collaborator Loading…
3 tasks
1
[CK] [FlyDSL] [Perf] split-K for the FlyDSL a8w8 preshuffle GEMM
#4908 opened Aug 21, 2026 by XiaobingSuper Contributor Loading…
7
[FlyDSL] Add GDN MTP kernels: causal conv1d update and gated delta rule
#4907 opened Aug 21, 2026 by yiijin Contributor Loading…
1 task
6
[FlyDSL] update gfx950 layout fmoe2 csv
#4906 opened Aug 21, 2026 by charlieguo1106 Contributor Loading…
1 task
1
[FlyDSL] [Build] Use upstream parallel scheduler for AOT builds
#4905 opened Aug 21, 2026 by zhiding512 Contributor Loading…
3 tasks done
2
[OPUS] [JIT] [Build] opus bf16 gemm co integration
#4903 opened Aug 21, 2026 by demonsan Contributor Loading…
1 task
1
[MLA PS] workload-aware KV-split count for auto mode
#4899 opened Aug 21, 2026 by zejunchen-zejun Contributor • Draft
1
[Triton/Gluon] [CI] [DO NOT MERGE] Triton release test ci:atom_full
#4897 opened Aug 21, 2026 by yuyzhang512 Contributor Loading…
1 task
1
[Triton/Gluon] for k3 sa submit
#4896 opened Aug 21, 2026 by gbyu-amd Contributor • Draft
1 task
1
[Triton/Gluon] Add MXFP4/MXFP8 gfx950/gfx1250 gluon quant kernels
#4894 opened Aug 21, 2026 by NimitPtl Contributor Loading…
1 task
10
[ASM] 1x32 mxfp4 asm kernel
#4890 opened Aug 20, 2026 by JohnNikolay84 Contributor Loading…
1 task
@JohnNikolay84
1
remove lean_atten files and relevant imports in aiter ci:triton-300x ci:triton-355
#4887 opened Aug 20, 2026 by nidal567 Contributor Loading…
1 task done
@nidal567
1
[FlyDSL] feat(flydsl): Add HSTU Backward kernel
#4885 opened Aug 20, 2026 by SamiAario-AMD • 2/2 Loading…
1 task done
27
[FlyDSL] Fused K5 + K6 gfx942 kernel for linear attention prefill
#4884 opened Aug 20, 2026 by vpietila-amd Contributor • Draft
1 task done
25
[Triton/Gluon] [QSA] Add paged sparse attention kernels
#4882 opened Aug 20, 2026 by haic0 Loading…
1
[CI] ci: add DCO signoff check
#4881 opened Aug 20, 2026 by gyohuangxin Member • Draft
3
[Triton/Gluon] fix pa_prefill perf regression for small HEAD_DIM ci:triton-300x
#4880 opened Aug 20, 2026 by mengfei-jiang Contributor Loading…
1 task done
1
[Triton/Gluon] triton fp4gemm disable gluon on gfx1250
#4878 opened Aug 20, 2026 by junhaha666 Contributor Loading…
1
[HIP] [GFX950]OPUS MHA Support LSE
#4877 opened Aug 20, 2026 by shay-li77 Contributor Loading…
1 task
6
[FlyDSL] remove flydsl_moe2 v1 code
#4876 opened Aug 20, 2026 by charlieguo1106 Contributor Loading…
1 task
1
ProTip! Find all pull requests that aren't related to any open issues with -linked:issue.