Skip to content

Pull requests: ROCm/aiter

New pull request New
214 Open 2,209 Closed
214 Open 2,209 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] Adapt Model Benchmarking Scripts to the New bench_mha.py CLI bug Something isn't working ci:triton-355 triton
#2673 opened Apr 9, 2026 by brunomazzottiamd Contributor Loading…
1 task done
@brunomazzottiamd
1
Add ROPE to unified attention
#2672 opened Apr 9, 2026 by anhminhnguyenhoang Contributor • Draft
1 task
Alizaidy/issue 2656 ci:triton-355
#2671 opened Apr 9, 2026 by azaidy Contributor Loading…
Add release engineering infrastructure
#2670 opened Apr 9, 2026 by sunway513 Collaborator Loading…
5 tasks
8
[TRITON] Reduce GMM tests ci:triton-355 enhancement New feature or request triton
#2669 opened Apr 9, 2026 by brunomazzottiamd Contributor Loading…
1 task done
@brunomazzottiamd
2
gemm tunner: add baseline path and only save config when better than baseline path
#2668 opened Apr 9, 2026 by XiaobingSuper Contributor Loading…
1 task
3
CI: Enable Deepseek ATOM tests on MI35X ci:atom
#2667 opened Apr 9, 2026 by gyohuangxin Member Loading…
2
fix(setup.py): accept FlyDSL dev/rc builds when version matches pinned release
#2664 opened Apr 9, 2026 by guangzlu Contributor Loading…
3
[TRITON] Prevent NUM_KSPLIT from reducing K dim below GROUP_K ci:triton-355
#2661 opened Apr 9, 2026 by vgokhale Contributor Loading…
2
[Perf]support ck fp8 blockscale splitk in moe tunner
#2658 opened Apr 8, 2026 by lalala-sh Contributor Loading…
1 task done
2
add triton fallback for mi455 gptoss & dsfp4
#2657 opened Apr 8, 2026 by HaonanWang98 Loading…
1 task
3
[FIX] fix ck tile shim perf eval in mha bwd v3
#2655 opened Apr 8, 2026 by JaxChen29 Contributor Loading…
1 task
2
[OPUS] tiled mma with step_k and dword copy by set_slice_impl
#2652 opened Apr 8, 2026 by kaiyang-1 Contributor • Draft
3
feat: add flydsl moe aot
#2651 opened Apr 8, 2026 by zhiding512 Contributor Loading…
6
Fuse qk norm cache group quant
#2649 opened Apr 8, 2026 by yzhou103 Contributor Loading…
1 task
8
feat: _flash_attn_forward add out args
#2648 opened Apr 8, 2026 by RuibinCheung Loading…
1 task done
4
refactor_hip_kernel
#2647 opened Apr 8, 2026 by amd-ruitang3 Contributor Loading…
1 task
16
[TRITON] mHC/mHC-lite: Manifold-constrained Hyper Connection
#2646 opened Apr 7, 2026 by anhminhnguyenhoang Contributor • Draft
1 task
1
fix(ck_gemm): fix multi-arch build targeting and kernel dispatch across all CK GEMM modules ci:sglang ci:vllm
#2645 opened Apr 7, 2026 by eppaneamd Loading…
26
Enable Grouped-Query Attention (GQA) based on MHA
#2643 opened Apr 7, 2026 by etemadiamd Contributor Loading…
1 task
1
fix: enable MXFP4 MoE at TP=4/8 via CKTile a4w4 kernels and quant fixes ci:all
#2642 opened Apr 7, 2026 by thpereir Loading…
1 task
2
Restore CKTile MOE tuning and blockPerCu support in MoE tuner
#2640 opened Apr 7, 2026 by amd-yashagar Contributor Loading…
7
Annotate moe_sorting_dispatch_policy as int for fused_moe
#2639 opened Apr 7, 2026 by nholmber Loading…
1 task done
1
1
[config] Add bf16 tuned GEMM config for Kimi-K2.5 on MI355 (TP8)
#2632 opened Apr 7, 2026 by akao-amd Loading…
1 task
Add PA_PS 8-wave kernel for MI308 with co-execution
#2630 opened Apr 7, 2026 by quintinwang5 Loading…
1 task
1
ProTip! no:milestone will show everything without a milestone.