Skip to content

Pull requests: ROCm/hip-ep

Pull request creation is restricted
35 Open 562 Closed
35 Open 562 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

perf(linear_attention): halve the prefill chunk and split the scan over dv (-25% op, -1.07 s TTFT, 129 -> 49 MiB scratch)
#627 opened Aug 2, 2026 by BoarQing Contributor • Draft
5
fix(matmul-lowering): use extractMemRefDataPtr for offset-correct pointers
#626 opened Aug 1, 2026 by kirankumarmatam-amd Collaborator • Draft
3
fix(promote-strided): include hip.npi_* ops in strided-input promotion
#625 opened Aug 1, 2026 by kirankumarmatam-amd Collaborator • Draft
3
fix(softmax): derive elem_size_bytes from MemRef type; add fp32 softmax kernel
#624 opened Aug 1, 2026 by kirankumarmatam-amd Collaborator • Draft
3
perf(gqa): template the sliding window out of the no-window v5 prefill path (-494 ms TTFT)
#623 opened Aug 1, 2026 by BoarQing Contributor Loading…
4 of 5 tasks
5
Fix GQA default scale for packed QKV
#621 opened Jul 31, 2026 by Vandana9628 Contributor Loading…
3
fix(runtime): defer MIOpen handle fetch for vendor-BLAS-disabled elem…
#619 opened Jul 31, 2026 by amd-mah Loading…
1 of 4 tasks
3
feat(hipsr): add hipsr.compute and hipsr.compute_yield ops
#615 opened Jul 31, 2026 by zpye Collaborator Loading…
4
refactor(hipsr): move shape regions to placeholders large-pr Soft size threshold (>500 LOC or >10 files). Reviewer signal, not a block.
#610 opened Jul 31, 2026 by zpye Collaborator • Draft
3
fix(gbq): key GatherBlockQuantized default zero point off ONNX storage type, not signedness
#609 opened Jul 31, 2026 by BoarQing Contributor • Draft
4 tasks
3
feat(onnx-to-hip): stamp Slice params before constant externalization
#607 opened Jul 31, 2026 by amd-mounikk Contributor Loading…
1 of 3 tasks
1
feat(runtime): add fp32 support to ReduceSum kernel
#604 opened Jul 31, 2026 by AMDmoore Contributor Loading…
2 tasks
3
Perf/pr596 linear attention
#603 opened Jul 31, 2026 by donghanbj Contributor • Draft
4 tasks
5
HIP EP Enablement for MI350 enhancement New feature or request
#602 opened Jul 31, 2026 by penglu12-quark Loading…
@penglu12-quark
3
perf(qmoe): group prefill expert matmuls
#601 opened Jul 30, 2026 by donghanbj Contributor • Draft
4 tasks
5
perf(kernels): speed up linear-attention prefill ~2.6x (Q/K LDS tiling + dk-split)
#600 opened Jul 30, 2026 by donghanbj Contributor Loading…
3 of 4 tasks
5
test: update OGA version to supports features of gemma4-unified.
#599 opened Jul 30, 2026 by amd-genmingz Collaborator • Draft
3
fix(morphizen): guard empty output NodeArg in try_fuse
#594 opened Jul 30, 2026 by AMDmoore Contributor Loading…
3 of 4 tasks
7
feat(hipsr): add expand LLVM lowering
#585 opened Jul 29, 2026 by mingyueliuh Collaborator • Draft
2 tasks done
3
fix(hip): unify result shapes and accept dynamic batched MatMul large-pr Soft size threshold (>500 LOC or >10 files). Reviewer signal, not a block.
#581 opened Jul 28, 2026 by fhanuman Collaborator • Draft
3
ci(winml): standalone MSIX packaging workflow, ROCm runtime 7.2.2609.47 (do not merge)
#579 opened Jul 28, 2026 by zz002 Collaborator • Draft
7 of 9 tasks
3
fix(typeid): add explicit TypeIDs to all anonymous-namespace RewritePattern subclasses large-pr Soft size threshold (>500 LOC or >10 files). Reviewer signal, not a block.
#573 opened Jul 28, 2026 by kirankumarmatam-amd Collaborator • Draft
5 tasks done
4
feat(hipsr): lifetime-based pool allocation (single group)
#572 opened Jul 28, 2026 by zz002 Collaborator • Draft
3 of 6 tasks
3
feat(compiler): pack dynamic pool allocs by lifetime (hip-pool-allocs)
#570 opened Jul 27, 2026 by fhanuman Collaborator Loading…
7
fix(use-output-allocator): report graph-output rank for reshaped-view outputs
#557 opened Jul 24, 2026 by anilmartha Contributor • Draft
1 of 3 tasks
4
ProTip! Add no:assignee to see everything that’s not assigned.