Skip to content

Pull requests: uxlfoundation/oneDNN

New pull request New
122 Open 3,491 Closed
122 Open 3,491 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

graph: backend: dnnl: exec: gated_mlp uses unique_ptr instead of raw ptr component:graph-api Codeowner: @oneapi-src/onednn-graph
#5060 opened Apr 21, 2026 by TaoLv Contributor Loading…
[WIP] xe: gated_mlp: improve performance of ukernel-based gmlp component:common component:tests Codeowner: @oneapi-src/onednn-arch platform:gpu-intel Codeowner: @oneapi-src/onednn-gpu-intel
#5059 opened Apr 21, 2026 by hidefromkgb Contributor • Draft
WIP: [GPU] Improve int8 ResNet-34 convolution performance on BMG platform:gpu-intel Codeowner: @oneapi-src/onednn-gpu-intel
#5058 opened Apr 20, 2026 by echeresh Contributor Loading…
1
[GPU][NVL-P] Add acc mode restriction on atomic acc platform:gpu-intel Codeowner: @oneapi-src/onednn-gpu-intel
#5057 opened Apr 20, 2026 by kealan-barbieri Contributor Loading…
4 tasks
4
[GPU] xe: tensor: fix view_t::normalized_tlayout() backport component:tests Codeowner: @oneapi-src/onednn-arch platform:gpu-intel Codeowner: @oneapi-src/onednn-gpu-intel
#5055 opened Apr 20, 2026 by echeresh Contributor Loading…
1
benchdnn: support SDPA primitive in verbose output component:tests Codeowner: @oneapi-src/onednn-arch documentation A request to change/fix/improve the documentation. Codeowner: @oneapi-src/onednn-doc
#5054 opened Apr 20, 2026 by kwieloch-intel Contributor • Draft
graph: fix a few moderate/minor coverity issues component:graph-api Codeowner: @oneapi-src/onednn-graph component:tests Codeowner: @oneapi-src/onednn-arch
#5051 opened Apr 20, 2026 by TaoLv Contributor Loading…
1
benchdnn: inputs: graph: add torchbench sdpa training tests component:tests Codeowner: @oneapi-src/onednn-arch
#5050 opened Apr 20, 2026 by TaoLv Contributor Loading…
1 of 3 tasks
1
[GPU] ngen: emul: fix s0Q*s1W multiply on XE3P. Backport backport third_party
#5048 opened Apr 17, 2026 by skazakov1 Contributor Loading…
3
[GPU] ngen: emul: fix s0Q*s1W multiply on XE3P third_party
#5046 opened Apr 17, 2026 by skazakov1 Contributor Loading…
9
WIP: xe: conv: consolidate accumulator type setup platform:gpu-intel Codeowner: @oneapi-src/onednn-gpu-intel
#5045 opened Apr 17, 2026 by echeresh Contributor Loading…
1
git: ignore .gdbinit and .gdb_history
#5044 opened Apr 17, 2026 by Sqvid Contributor Loading…
6
WIP: [GPU] Use f32 accumulator for f16 input in convolution component:tests Codeowner: @oneapi-src/onednn-arch platform:gpu-intel Codeowner: @oneapi-src/onednn-gpu-intel
#5041 opened Apr 16, 2026 by echeresh Contributor Loading…
4
cpu: x64: matmul: enable treat_as_plain for weights format bug A confirmed library bug platform:cpu-x64 Intel64/AMD64 processors. Codeowner: @oneapi-src/onednn-cpu-x64
#5038 opened Apr 16, 2026 by xuxinzen Contributor Loading…
8
cpu: aarch64: make jit_uni_pool SVE instantiation vector length agnostic component:common platform:cpu-aarch64 Codeowner: @oneapi-src/onednn-cpu-aarch64
#5036 opened Apr 16, 2026 by Sqvid Contributor Loading…
3 tasks done
1
Test PR. Upconvert fp8 weights to xf16 in Matmul in case of xf16 activations for 3.10 backport component:common component:tests Codeowner: @oneapi-src/onednn-arch platform:cpu-x64 Intel64/AMD64 processors. Codeowner: @oneapi-src/onednn-cpu-x64
#5035 opened Apr 16, 2026 by asimonov1 Contributor • Draft
2
xe: gemm: use dispatch table for simple strategy parameter parsing platform:gpu-intel Codeowner: @oneapi-src/onednn-gpu-intel
#5025 opened Apr 15, 2026 by Simonsays095 Contributor Loading…
3
[GPU] GEMM Acc fixup platform:gpu-intel Codeowner: @oneapi-src/onednn-gpu-intel
#5017 opened Apr 14, 2026 by kealan-barbieri Contributor Loading…
2 of 4 tasks
7
cpu: x64: matmul: Enable int8 grouped quantization platform:cpu-x64 Intel64/AMD64 processors. Codeowner: @oneapi-src/onednn-cpu-x64
#5014 opened Apr 14, 2026 by inteldimitrius Contributor • Draft
6
cpu: aarch64: enable binary op and binary post-ops on ASIMD platform:cpu-aarch64 Codeowner: @oneapi-src/onednn-cpu-aarch64
#5008 opened Apr 13, 2026 by renato-arantes Contributor Loading…
3 tasks done
13
oneDNN v3.12 release notes backport documentation A request to change/fix/improve the documentation. Codeowner: @oneapi-src/onednn-doc
#5005 opened Apr 11, 2026 by vpirogov Contributor Loading…
9
aarch64: support for per_dim_0 scales and bf16 dst_dt in jit int8 matmul component:common component:tests Codeowner: @oneapi-src/onednn-arch platform:cpu-aarch64 Codeowner: @oneapi-src/onednn-cpu-aarch64
#4987 opened Apr 9, 2026 by michalowski-arm Contributor Loading…
2 tasks done
7
MFDNN-14690: Replace XE3P_35_10/11/UNKNOWN Core enum values with Xe3p platform:gpu-intel Codeowner: @oneapi-src/onednn-gpu-intel third_party
#4981 opened Apr 8, 2026 by dyoussif Contributor Loading…
39
cpu, benchdnn: add reorder to/from grouped with different dts and use grouped in matmul ref component:tests Codeowner: @oneapi-src/onednn-arch
#4977 opened Apr 8, 2026 by mzhukova Contributor • Draft
@mzhukova
1
cpu: aarch64: implement forward lnorm in SVE component:common platform:cpu-aarch64 Codeowner: @oneapi-src/onednn-cpu-aarch64
#4974 opened Apr 8, 2026 by Sqvid Contributor • Draft
3 tasks done
7
ProTip! What’s not been updated in a month: updated:<2026-03-20.