1.24.0 release cherry-pick round 1#27103

Closed

tianleiwu wants to merge 28 commits intomainfrom

tlwu/rel-1.24.0_cherrypick_round_1

Commits on Jan 21, 2026

Add type definitions, registration, utilities for INT2/UINT2 support (#26824)

vraspar
authored and
tianleiwu
committed
[QNN EP] Add BFloat16 dtype support in QNN EP (#26987)

authored and
tianleiwu
committed
Implement new experimental lookup-based matrix multiplication method(TMAC) (#26695)

authored and
tianleiwu
committed
[MLAS/CPU EP] Improve performance of Silu activation path within the QuickGelu CPU kernel (#26753)

authored and
tianleiwu
committed
[QNN EP] Add support for handling 0-dimension for Concat Op (#27000)

qti-ashwshan
authored and
tianleiwu
committed
Fix ClipQuantFusion crash when Clip has multiple input edges (#27016)

authored and
tianleiwu
committed
[QNN EP] Support quantized BatchNorm with per-channel DQ params on QNN HTP (#26959)

qti-yuduo
authored and
tianleiwu
committed
Add API to get ep graph partitioning info (#26781)

authored and
tianleiwu
committed
[OVEP] OpenVINO EP Features and bug-fixes for ORT-1.24 - Follow up (#27007)

committed
[QNN-EP] Add MatMulNBits translation for GPU (#26340)

authored and
tianleiwu
committed
[MLAS/NEON] Add dedicated kernel for depthwise convolution for ARM64 using NEON intrinsics (#26688)

authored and
tianleiwu
committed
[QNN-EP] Support alternate Layernorm fusion pattern in QNN preprocess (#26060)

qti-mattsinc
authored and
tianleiwu
committed
Implement multithreading in qgemm_kleidi (#26301)

authored and
tianleiwu
committed
[CXX] Enable users to specify custom OrtSyncStream via RunOptions (#26988)

yuslepukhin
authored and
tianleiwu
committed
Added support for QMX kernels in MLAS (#26849)

qti-vaiskv
authored and
tianleiwu
committed
Tweak external resource importer handle structs (#27040)

skottmckay
authored and
tianleiwu
committed
[QNN EP] Add QuickGELU operator support for QNN provider (#27034)

authored and
tianleiwu
committed
Add INT2 and UINT2 support for QDQ, transpose and cast ops (#27022)

vraspar
authored and
tianleiwu
committed
Introducing BF16 Pointwise NCHWc Convolution for Arm64 (#26838)

Rohanjames1997
authored and
tianleiwu
committed
[EP ABI] Add CreateCustomOpDomains() API for plugin EP to register custom ops (#27050)

chilo-ms
authored and
tianleiwu
committed
Add a new pipeline for CUDA 13 nuget builds (#27023)

authored and
tianleiwu
committed
[EP ABI] Update Graph_GetGraphView() implementation (#26711)

chilo-ms
authored and
tianleiwu
committed
[webgpu] Fix a bug for im2col (#27069)

wenqinI
authored and
tianleiwu
committed
[QNN EP] Add FusedMatMul operator support (#27044)

tirupath-qti
authored and
tianleiwu
committed
Disable Float32_2Bits_Asymmetric_256x256 test (#27046)

vraspar
authored and
tianleiwu
committed
Fix Doxygen documentation build error in onnxruntime_c_api.h (#27083)

nieubank
authored and
tianleiwu
committed
Print tensor for new packed type of 2 bits (#27064)
tianleiwu
committed
Fix GPU JAR testing on Linux (#27011)

authored and
tianleiwu
committed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

1.24.0 release cherry-pick round 1#27103

1.24.0 release cherry-pick round 1#27103
tianleiwu wants to merge 28 commits intomainfrom
tlwu/rel-1.24.0_cherrypick_round_1

Commits on Jan 21, 2026

Add type definitions, registration, utilities for INT2/UINT2 support (#26824)

[QNN EP] Add BFloat16 dtype support in QNN EP (#26987)

Implement new experimental lookup-based matrix multiplication method(TMAC) (#26695)

[MLAS/CPU EP] Improve performance of Silu activation path within the QuickGelu CPU kernel (#26753)

[QNN EP] Add support for handling 0-dimension for Concat Op (#27000)

Fix ClipQuantFusion crash when Clip has multiple input edges (#27016)

[QNN EP] Support quantized BatchNorm with per-channel DQ params on QNN HTP (#26959)

Add API to get ep graph partitioning info (#26781)

[OVEP] OpenVINO EP Features and bug-fixes for ORT-1.24 - Follow up (#27007)

[QNN-EP] Add MatMulNBits translation for GPU (#26340)

[MLAS/NEON] Add dedicated kernel for depthwise convolution for ARM64 using NEON intrinsics (#26688)

[QNN-EP] Support alternate Layernorm fusion pattern in QNN preprocess (#26060)

Implement multithreading in qgemm_kleidi (#26301)

[CXX] Enable users to specify custom OrtSyncStream via RunOptions (#26988)

Added support for QMX kernels in MLAS (#26849)

Tweak external resource importer handle structs (#27040)

[QNN EP] Add QuickGELU operator support for QNN provider (#27034)

Add INT2 and UINT2 support for QDQ, transpose and cast ops (#27022)

Introducing BF16 Pointwise NCHWc Convolution for Arm64 (#26838)

[EP ABI] Add CreateCustomOpDomains() API for plugin EP to register custom ops (#27050)

Add a new pipeline for CUDA 13 nuget builds (#27023)

[EP ABI] Update Graph_GetGraphView() implementation (#26711)

[webgpu] Fix a bug for im2col (#27069)

[QNN EP] Add FusedMatMul operator support (#27044)

Disable Float32_2Bits_Asymmetric_256x256 test (#27046)

Fix Doxygen documentation build error in onnxruntime_c_api.h (#27083)

Print tensor for new packed type of 2 bits (#27064)

Fix GPU JAR testing on Linux (#27011)