Skip to content

ci(TRI-1406): prepare ONNX Runtime backend for 26.06 (TRT 11, OV 2026.2, CCCL fix) (cherry-pick) - #345

Merged
mc-nv merged 1 commit into
mainfrom
mchornyi/TRI-1431/advance-default-branch
Jun 29, 2026
Merged

ci(TRI-1406): prepare ONNX Runtime backend for 26.06 (TRT 11, OV 2026.2, CCCL fix) (cherry-pick)#345
mc-nv merged 1 commit into
mainfrom
mchornyi/TRI-1431/advance-default-branch

Conversation

@mc-nv

@mc-nv mc-nv commented Jun 27, 2026

Copy link
Copy Markdown
Contributor

What does the PR do?

Cherry-picks 7db2534 from r26.06 to advance main post-26.05 toward the 26.07dev_user container train (tracks upstream 26.06).

Brings forward 26.06-cycle changes to tools/gen_ort_dockerfile.py: OpenVINO 2026.2, TensorRT 11, and the ORT #28972 CCCL header-ordering fix.

Related PRs:

Test plan:

  • gen_ort_dockerfile.py renders the expected Dockerfile against CUDA 13.3 / TRT 11.
  • CI Pipeline ID: TBD

Related Issues:

  • Resolves: TRI-1431

….2, CCCL fix) (#344)

* ci(TRI-1406): add OpenVINO 2026.2

* ci(TRI-1406): picking support for TensorRT 11

* including patch for ORT

* ci(TRI-1406): cherry-pick ORT PR #28972 for CCCL header-ordering fix

Adds a third cherry-pick on top of the existing ORT 1.26.0 patch chain
to fix a cudafe++ parse error in <cub/device/device_transform.cuh:45>
seen under CUDA 13.3 (`global qualification of class name is invalid
before ':' token` at the `cuda::proclaims_copyable_arguments`
specialization).

Upstream draft PR: microsoft/onnxruntime#28972

* remove 26.06 related changes.
@mc-nv
mc-nv merged commit bb4926b into main Jun 29, 2026
3 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants