Skip to content

Pull requests: ROCm/aiter

New pull request New
103 Open 950 Closed
103 Open 950 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

CI: Move triton and sglang from mi325 to mi355
#1183 opened Oct 13, 2025 by gyohuangxin Loading…
1
[CK_TILE] FMHA BWD Optimizations for D48 for GFX950
#1180 opened Oct 13, 2025 by DDEle Loading…
1 task done
1
[MI35X] Enhance mha bwd varlen kernels
#1179 opened Oct 13, 2025 by slippedJim Loading…
1 task
CI: Temporarily using mi355 runners in Triton Tests util issues on mi…
#1178 opened Oct 13, 2025 by gyohuangxin Loading…
1
fix torch compile when using fp8 fla
#1177 opened Oct 13, 2025 by guangzlu Loading…
1 task
2
[Feat] Deepgemm FP8 paged_mqa_logits
#1174 opened Oct 13, 2025 by sjfeng1999 Loading…
1 task done
4
Add triton_metadata_redirect module with with_metadata_path decorator
#1172 opened Oct 13, 2025 by jwu10003 Loading…
1 task
add hipblaslt swizzle feature
#1169 opened Oct 11, 2025 by yixionghuo Loading…
1 task
add define
#1168 opened Oct 11, 2025 by yixionghuo Loading…
1 task
refactor mha fwd and bwd args
#1165 opened Oct 11, 2025 by minmengdie Loading…
1 task
23
CI: Operators tunning pipelines
#1163 opened Oct 11, 2025 by gyohuangxin • Draft
[Mi35x] A8w8 asm codegen and tune
#1161 opened Oct 11, 2025 by yzhou103 Loading…
1 task
5
[355_wip_triton] [triton] fuse bf16_gemm_reduce_kernel + rope_kv_cache
#1153 opened Oct 9, 2025 by k50112113 Loading…
[355_wip] [triton] fuse bf16_gemm_reduce_kernel + rope_kv_cache
#1151 opened Oct 9, 2025 by k50112113 Loading…
Catchall PR for all 355_wip related changes
#1148 opened Oct 9, 2025 by azaidy • Draft
1 task
[BugFix] Change Vskip Selection Logic invalid This doesn't seem right
#1136 opened Oct 8, 2025 by vllmellm Loading…
1 task
1
1
fmha v3 API only generate for supported targeting GPU arch
#1134 opened Oct 7, 2025 by HollowMan6 Loading…
1 task done
16
Various benchmark script fixes
#1133 opened Oct 7, 2025 by jayfurmanek • Draft
Bump composable_kernel for fmha 192x128 head dims commit
#1130 opened Oct 6, 2025 by samremes • Draft
1 task
[CK_TILE] Temporarily disable k length=1 test cases in seqence padding
#1129 opened Oct 3, 2025 by Jeff-Huang Loading…
[Triton] e2e fused MoE for small N and fp8 blockscale MoE benching
#1126 opened Oct 2, 2025 by juuso-oskari Loading…
@juuso-oskari @Chi-Chu319
5
update sparse mla kernel
#1125 opened Oct 2, 2025 by Zzz9990 • Draft
1 task
lean_attention: add GQA support across kernel and wrapper; add tests
#1123 opened Oct 1, 2025 by kesavanramakrishnan Loading…
3
[Triton] FP4 GEMM weight shuffle support and tunning
#1120 opened Sep 30, 2025 by k50112113 Loading…
Mla splitkv enhance split alg inte
#1108 opened Sep 28, 2025 by fangche123 Loading…
1 task
3
ProTip! Find all pull requests that aren't related to any open issues with -linked:issue.