Skip to content

Pull requests: ROCm/aiter

New pull request New
347 Open 2,765 Closed
347 Open 2,765 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

re-tune flydsl bf16 gemm tuned config
#3434 opened May 30, 2026 by XiaobingSuper Contributor Loading…
1 task
2
Align aiter prebuild jobs with build runner CPU limit
#3433 opened May 30, 2026 by gyohuangxin Member Loading…
2
Fix FlyDSL bf16 GEMM tuned CSV kernel names
#3432 opened May 30, 2026 by coderfeli Collaborator Loading…
1
Add Kimi-K2 Thinking entry to per-model kernel benchmark sweep
#3431 opened May 30, 2026 by okorzh-amd Contributor Loading…
1
Add native integer all-gather dtype support and optimize gfx942 custom all-gather
#3430 opened May 29, 2026 by hubertlu-tw Contributor Loading…
5 of 6 tasks
2
Fuse dynamic_per_tensor_quant_fp8_i8 into one launch for the decode regime
#3429 opened May 29, 2026 by JohnQinAMD Contributor Loading…
2 of 3 tasks
5
add MX_FP4_A8 tuned configs and dispatch for moe_gemm_a8w4
#3428 opened May 29, 2026 by xiaohuguo2023 Member Loading…
21
[FLYDSL] Measure impact of new stream-k a16w16 gemm
#3427 opened May 29, 2026 by xytpai Contributor Loading…
2
Add bf16 qh8 qseqlen=2 MLA decode kernel (gfx950)
#3426 opened May 29, 2026 by alexioslyrakis-amd Contributor Loading…
3 tasks done
1
[triton-mha] hint head-stride div-by-8 for vectorized global load
#3424 opened May 29, 2026 by mgehre-amd Contributor • Draft
1
[triton-mha] add gfx1151 tuning config
#3423 opened May 29, 2026 by mgehre-amd Contributor • Draft
2 of 3 tasks
1
Add GLM-5.1 FP8 blockscale GEMM/FMoE tunings for gfx942 (MI300X/MI325)
#3422 opened May 29, 2026 by akii96 Contributor Loading…
2
[gfx1151] [flash_attn_triton_amd]: Tune for gfx1151 / ViT
#3419 opened May 29, 2026 by mgehre-amd Contributor • Draft
1
Add PER_TOKEN_HEAD FP8 quantization and P-scale for mha_batch_prefill
#3418 opened May 29, 2026 by msaffari-amd Loading…
1 task done
5
fmoe: Expected m host hint
#3414 opened May 29, 2026 by inkcherry Contributor Loading…
1 task
5
[Bugfix] Enable MXFP4 MoE at TP=4/8 via CKTile a4w4 kernels and quant fixes
#3412 opened May 29, 2026 by haoyangli0109 Contributor Loading…
2
[Bugfix]Fix varctx scheduling in pa_mqa_logits benchmark.
#3410 opened May 29, 2026 by charlieguo1106 Loading…
1
mla
#3409 opened May 29, 2026 by feifei14119 Contributor • Draft
1 task
1
[FlyDSL MoE] Add no combine feature
#3408 opened May 28, 2026 by zx3xyy Loading…
1 task done
6
[Triton] Add New Features and Performance Improvement for GMM Kernel ci:triton-300x ci:triton-355 enhancement New feature or request triton
#3407 opened May 28, 2026 by brunomazzottiamd Contributor Loading…
4 tasks done
@brunomazzottiamd
7
[Gluon] Moe a8w4 preshuffle weights
#3405 opened May 28, 2026 by nsusanto Contributor Loading…
1
[OPUS]Opus gemm 4G reject
#3403 opened May 28, 2026 by demonsan Contributor Loading…
1
[Triton-Gluon-MLA-GFX950] add mla_decode_gluon_bh16_dcp kernel wrapper for DCP MLA decode
#3402 opened May 28, 2026 by Dewei-Wang-sh Contributor Loading…
13
[FLYDSL MOE] fix MoE accuracy with openning pad
#3401 opened May 28, 2026 by Zzz9990 Contributor • Draft
1 task
1
Fix reduce scatter when dim != 0
#3395 opened May 28, 2026 by yitingw1 Loading…
1 task
2
ProTip! Type g i on any issue or pull request to go back to the issue listing page.