Skip to content

Pull requests: ROCm/aiter

New pull request New
378 Open 2,900 Closed
378 Open 2,900 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

Support padded K for a8w8 bpreshuffle GEMM
#3611 opened Jun 8, 2026 by ThomasNing Loading…
1
Zhimding/gfx1250 moe 2mode e2e v1 0608
#3610 opened Jun 8, 2026 by zhiding512 Contributor • Draft
1 task
Add GLM GQA FP8 KV paged attention test
#3609 opened Jun 8, 2026 by ThomasNing Loading…
2
Refine flydsl gemm config selection code
#3608 opened Jun 8, 2026 by xytpai Contributor Loading…
1
Use flyc.compile for MOE kernel fast dispatch
#3607 opened Jun 8, 2026 by xudoyuan Contributor Loading…
1 task
1
[Bugfix][MLA] Correct final_lse in PS MLA prefill kernel for chunked prefill
#3606 opened Jun 8, 2026 by simondanielsson • Draft
1 task
1
CI: add a16w16 GEMM to op tuning workflow
#3605 opened Jun 8, 2026 by gyohuangxin Member • Draft
3
fix deepgemm_fp8_paged_mqa_logits kernel input parameter mismatch issue
#3604 opened Jun 8, 2026 by lirui927 Loading…
1 task done
1
[Perf] Use fused silu_and_mul for CK-Tile interleaved post-activation
#3603 opened Jun 8, 2026 by yichiche Contributor Loading…
1
[FLYDSL] Optimize GDR prefill chunk_gdn_fwd_h kernel for MI35X
#3602 opened Jun 8, 2026 by huizzhan Contributor • Draft
1 task
1
[TRITON] Fix Assert in Triton fused_kv_cache
#3601 opened Jun 8, 2026 by leonling-ll Loading…
3
Update flydsl to 0.2.0.dev20260608+c957349 ci:all
#3600 opened Jun 8, 2026 by xudoyuan Contributor Loading…
1 task
1
[Feat] Support FP4 gather_kv_b_proj
#3597 opened Jun 8, 2026 by qichu-yun Loading…
1 task
1
CI: add FFM bringup workflow
#3596 opened Jun 8, 2026 by gyohuangxin Member • Draft
6
[module_causal_conv1d_update] refactor hip kernel
#3595 opened Jun 8, 2026 by amd-ruitang3 Contributor Loading…
1 task
10
[OPUS] gfx942 a16w16 bf16 GEMM splitK and WKC pipeline family
#3594 opened Jun 8, 2026 by yifehuan Contributor Loading…
1 task
1
MOE: add AITER_MOE_FORCE_BF16_ACT to force bf16 activations (opt-in)
#3593 opened Jun 8, 2026 by sphinx07 Loading…
1
[hotfix] always use fp4x2 for swiglu separated per_1x32 path
#3591 opened Jun 8, 2026 by yadaish Contributor Loading…
1
[op_tests] Refactor MoE legacy UT into per-quant smoke sweep
#3585 opened Jun 7, 2026 by zhiding512 Contributor Loading…
4
[feat] FP8 (DeepSeek-V4 layout) sparse paged prefill attention
#3583 opened Jun 7, 2026 by carlushuang Collaborator Loading…
2 of 3 tasks
1
[gfx1250][FlyDSL] Add Ragged-M OOB Support for PTPC FP8 GEMM
#3582 opened Jun 7, 2026 by aoli26 Contributor Loading…
1 task done
18
mha_native: native HIP D64 BF16 split-K forward backend for flash_attn_func
#3581 opened Jun 6, 2026 by rocking5566 Contributor Loading…
2 of 3 tasks
7
gfx950 MoE A8W4: tuned entries for gpt-oss shapes + fallback hardening
#3580 opened Jun 6, 2026 by xiaohuguo2023 Member Loading…
2
ci: add paired-release validation gate workflow (AITER+ATOM matrix)
#3578 opened Jun 6, 2026 by sunway513 Collaborator Loading…
2 of 3 tasks
6
[TRITON][GLUON] Unified attention 2d gluon kernel
#3577 opened Jun 6, 2026 by cagrikymk Contributor Loading…
1
ProTip! Add no:assignee to see everything that’s not assigned.