ORT 1.28.0 release cherry-pick round 1 (#29771)
This cherry-picks the following commits for the release:
| Commit ID | PR Number | Commit Title |
|-----------|-----------|-------------|
| dd32f3598a | #29590 | Fix libcudart.so.13 hard dependency in pybind
module breaking import on CPU-only Linux |
| cc44a4ddde | #29706 | [CUDA] Fix XQA GroupQueryAttention
cudaErrorInvalidValue on Blackwell (sm_120) |
| 23a7e9de77 | #29705 | [CUDA] Do not link nvrtc |
| ee93f8332a | #29711 | [CUDA] Update cuda arch list for packages of
cuda 12.8 |
| fea45a315f | #29620 | [CUDA] Add cuDNN-free ArgMax/ArgMin/ReduceSum
and fix LogSoftmax on plugin EP |
| f05b21861c | #29624 | Enable Spectre-mitigated MSVC libs for BinSkim
builds |
| 1c89b863c9 | #29687 | [BUILD] CUDA_QUANT_PREPROCESS off by default and
Adjust CI |
| 41bd391176 | #29658 | [CUDA] Fix null allocator passed to plugin EP
kernel PrePack |
| 405fbea350 | #28896 | Add Windows ARM64 CUDA plugin package and align
CUDA metadata/artifact naming |
| 308f24c36c | #29622 | Enable fpA_intB GEMM in CUDA builds and add
configurable options |
| 16ebc1d98b | #29731 | [Build] Use GPU pool to unblock CI temporarily |
|5911a3a263| #29748 | Add OrtErrorCode::ORT_DEVICE_RESET |
|6217f73ec5 | #29663 | Fix plugin EP allocator deleter lifetime |
---------
Co-authored-by: Copilot <198982749+Copilot@users.noreply.github.com>
Co-authored-by: GitHub Copilot <copilot@example.com>
Co-authored-by: Edward Chen <18449977+edgchen1@users.noreply.github.com>
Co-authored-by: Yen-Shi Wang <yenshiw@nvidia.com>
