Skip to content

[None][feat] Upgrade NIXL to v1.0.1 and UCX to 1.21 - #14436

Merged
Shixiaowei02 merged 6 commits into
NVIDIA:mainfrom
chuangz0:upgrade_nixl_010
Jun 1, 2026
Merged

[None][feat] Upgrade NIXL to v1.0.1 and UCX to 1.21#14436
Shixiaowei02 merged 6 commits into
NVIDIA:mainfrom
chuangz0:upgrade_nixl_010

Conversation

@chuangz0

@chuangz0 chuangz0 commented May 22, 2026

Copy link
Copy Markdown
Collaborator

Squashed from upstream PR #13841 (NVIDIA/TensorRT-LLM):

  • Upgrade NIXL to v1.0.1
  • Upgrade UCX to 1.21
  • install_ucx.sh: replace pre-existing UCX with symlink
  • Add lib/ucx into libenv for CI PackageSanityCheck
  • Update images
  • Fix transfer bug

Summary by CodeRabbit

  • Chores

    • Updated NIXL dependency to v1.0.1.
    • Updated UCX library to v1.21.x.
    • Refreshed Docker image tags for LLM builds.
    • Improved build pipeline environment setup for library resolution.
  • Tests

    • Adjusted memory management and cache manager test configurations.

Review Change Stack

Description

Test Coverage

PR Checklist

Please review the following before submitting your PR:

  • PR description clearly explains what and why. If using CodeRabbit's summary, please make sure it makes sense.

  • PR Follows TRT-LLM CODING GUIDELINES to the best of your knowledge.

  • Test cases are provided for new code paths (see test instructions)

  • If PR introduces API changes, an appropriate PR label is added - either api-compatible or api-breaking. For api-breaking, include BREAKING in the PR title.

  • Any new dependencies have been scanned for license and vulnerabilities

  • CODEOWNERS updated if ownership changes

  • Documentation updated as needed

  • Update tava architecture diagram if there is a significant design change in PR.

  • The reviewers assigned automatically/manually are appropriate for the PR.

  • Please check this after reviewing the above items as appropriate for this PR.

GitHub Bot Help

To see a list of available CI bot commands, please comment /bot help.

@chuangz0

Copy link
Copy Markdown
Collaborator Author

/bot run --stage-list "Build-Docker-Images"

@tensorrt-cicd

Copy link
Copy Markdown
Collaborator

PR_Github #49828 [ run ] triggered by Bot. Commit: d2f8edc Link to invocation

@chuangz0

Copy link
Copy Markdown
Collaborator Author

/bot run --disable-fail-fast

@tensorrt-cicd

Copy link
Copy Markdown
Collaborator

PR_Github #49878 [ run ] triggered by Bot. Commit: bc7fa6a Link to invocation

@tensorrt-cicd

Copy link
Copy Markdown
Collaborator

PR_Github #49828 [ run ] completed with state ABORTED. Commit: d2f8edc

Link to invocation

@chuangz0
chuangz0 force-pushed the upgrade_nixl_010 branch from bc7fa6a to fe0819d Compare May 22, 2026 11:27
@chuangz0

Copy link
Copy Markdown
Collaborator Author

/bot run --disable-fail-fast

@tensorrt-cicd

Copy link
Copy Markdown
Collaborator

PR_Github #49928 [ run ] triggered by Bot. Commit: fe0819d Link to invocation

@tensorrt-cicd

Copy link
Copy Markdown
Collaborator

PR_Github #49878 [ run ] completed with state ABORTED. Commit: bc7fa6a

Link to invocation

@tensorrt-cicd

Copy link
Copy Markdown
Collaborator

PR_Github #49928 [ run ] completed with state SUCCESS. Commit: fe0819d
/LLM/main/L0_MergeRequest_PR pipeline #39505 completed with status: 'FAILURE'

CI Report

⚠️ Action Required:

  • Please check the failed tests and fix your PR
  • If you cannot view the failures, ask the CI triggerer to share details
  • Once fixed, request an NVIDIA team member to trigger CI again

CI Agent Failure Analysis

Link to invocation

@chuangz0
chuangz0 force-pushed the upgrade_nixl_010 branch from 1e35d2e to c313620 Compare May 25, 2026 03:23
@chuangz0

Copy link
Copy Markdown
Collaborator Author

/bot run --disable-fail-fast

@tensorrt-cicd

Copy link
Copy Markdown
Collaborator

PR_Github #50134 [ run ] triggered by Bot. Commit: c313620 Link to invocation

@chuangz0
chuangz0 force-pushed the upgrade_nixl_010 branch from 73bf439 to d349f77 Compare May 25, 2026 08:17
@chuangz0

Copy link
Copy Markdown
Collaborator Author

/bot run --disable-fail-fast

@tensorrt-cicd

Copy link
Copy Markdown
Collaborator

PR_Github #50181 [ run ] triggered by Bot. Commit: d349f77 Link to invocation

@tensorrt-cicd

Copy link
Copy Markdown
Collaborator

PR_Github #50134 [ run ] completed with state ABORTED. Commit: c313620

Link to invocation

@tensorrt-cicd

Copy link
Copy Markdown
Collaborator

PR_Github #50181 [ run ] completed with state SUCCESS. Commit: d349f77
/LLM/main/L0_MergeRequest_PR pipeline #39724 completed with status: 'FAILURE'

CI Report

⚠️ Action Required:

  • Please check the failed tests and fix your PR
  • If you cannot view the failures, ask the CI triggerer to share details
  • Once fixed, request an NVIDIA team member to trigger CI again

CI Agent Failure Analysis

Link to invocation

@chuangz0
chuangz0 force-pushed the upgrade_nixl_010 branch from fd3fd7d to b99939c Compare May 26, 2026 05:49
@chuangz0

Copy link
Copy Markdown
Collaborator Author

/bot run --disable-fail-fast

@tensorrt-cicd

Copy link
Copy Markdown
Collaborator

PR_Github #50290 [ run ] triggered by Bot. Commit: b99939c Link to invocation

@chuangz0
chuangz0 force-pushed the upgrade_nixl_010 branch from 53f9a8b to 1ac7b73 Compare May 26, 2026 12:32
@chuangz0

Copy link
Copy Markdown
Collaborator Author

/bot kill

@chuangz0
chuangz0 marked this pull request as ready for review May 26, 2026 12:33
@chuangz0
chuangz0 requested a review from a team May 26, 2026 12:33
@chuangz0
chuangz0 requested review from a team as code owners May 26, 2026 12:33
@chuangz0
chuangz0 requested review from a team and tburt-nv May 26, 2026 12:33
@chuangz0

Copy link
Copy Markdown
Collaborator Author

/bot run --disable-fail-fast

@tensorrt-cicd

Copy link
Copy Markdown
Collaborator

PR_Github #50416 [ run ] triggered by Bot. Commit: d86a652 Link to invocation

@tensorrt-cicd

Copy link
Copy Markdown
Collaborator

PR_Github #50416 [ run ] completed with state SUCCESS. Commit: d86a652
/LLM/main/L0_MergeRequest_PR pipeline #39940 completed with status: 'FAILURE'

CI Report

⚠️ Action Required:

  • Please check the failed tests and fix your PR
  • If you cannot view the failures, ask the CI triggerer to share details
  • Once fixed, request an NVIDIA team member to trigger CI again

CI Agent Failure Analysis

Link to invocation

@chuangz0
chuangz0 force-pushed the upgrade_nixl_010 branch from d86a652 to e5fd61b Compare May 27, 2026 09:12
@chuangz0

Copy link
Copy Markdown
Collaborator Author

/bot run --disable-fail-fast

@tensorrt-cicd

Copy link
Copy Markdown
Collaborator

PR_Github #50512 [ run ] triggered by Bot. Commit: e5fd61b Link to invocation

@tensorrt-cicd

Copy link
Copy Markdown
Collaborator

PR_Github #50512 [ run ] completed with state FAILURE. Commit: e5fd61b
/LLM/main/L0_MergeRequest_PR pipeline #40020 completed with status: 'FAILURE'

CI Report

⚠️ Action Required:

  • Please check the failed tests and fix your PR
  • If you cannot view the failures, ask the CI triggerer to share details
  • Once fixed, request an NVIDIA team member to trigger CI again

CI Agent Failure Analysis

Link to invocation

@chuangz0
chuangz0 force-pushed the upgrade_nixl_010 branch from e5fd61b to 8802bbf Compare May 28, 2026 01:22
@chuangz0

Copy link
Copy Markdown
Collaborator Author

/bot run --disable-fail-fast

@tensorrt-cicd

Copy link
Copy Markdown
Collaborator

PR_Github #50647 [ run ] triggered by Bot. Commit: 8802bbf Link to invocation

@tensorrt-cicd

Copy link
Copy Markdown
Collaborator

PR_Github #50647 [ run ] completed with state FAILURE. Commit: 8802bbf
/LLM/main/L0_MergeRequest_PR pipeline #40138 completed with status: 'FAILURE'

CI Report

⚠️ Action Required:

  • Please check the failed tests and fix your PR
  • If you cannot view the failures, ask the CI triggerer to share details
  • Once fixed, request an NVIDIA team member to trigger CI again

CI Agent Failure Analysis

Link to invocation

chuangz0 added 6 commits May 29, 2026 09:06
Squashed from upstream PR NVIDIA#13841 (NVIDIA/TensorRT-LLM):
- Upgrade NIXL to v1.0.1
- Upgrade UCX to 1.21
- install_ucx.sh: replace pre-existing UCX with symlink
- Add lib/ucx into libenv for CI PackageSanityCheck
- Update images
- Fix transfer bug

Signed-off-by: Chuang Zhu <111838961+chuangz0@users.noreply.github.com>
Signed-off-by: Chuang Zhu <111838961+chuangz0@users.noreply.github.com>
Signed-off-by: Chuang Zhu <111838961+chuangz0@users.noreply.github.com>
Signed-off-by: Chuang Zhu <111838961+chuangz0@users.noreply.github.com>
Signed-off-by: Chuang Zhu <111838961+chuangz0@users.noreply.github.com>
Signed-off-by: Chuang Zhu <111838961+chuangz0@users.noreply.github.com>
@chuangz0
chuangz0 force-pushed the upgrade_nixl_010 branch from 8802bbf to 9fec05e Compare May 29, 2026 01:06
@chuangz0

Copy link
Copy Markdown
Collaborator Author

/bot run --disable-fail-fast

@tensorrt-cicd

Copy link
Copy Markdown
Collaborator

PR_Github #50922 [ run ] triggered by Bot. Commit: 9fec05e Link to invocation

@tensorrt-cicd

Copy link
Copy Markdown
Collaborator

PR_Github #50922 [ run ] completed with state SUCCESS. Commit: 9fec05e
/LLM/main/L0_MergeRequest_PR pipeline #40382 completed with status: 'SUCCESS'
Pipeline passed with automatic retried tests. Check the rerun report for details.

CI Report

Link to invocation

@yuanjingx87 yuanjingx87 left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

approved on behalf of trt-llm-oss-compliance

@Shixiaowei02
Shixiaowei02 merged commit 2f9b85a into NVIDIA:main Jun 1, 2026
7 checks passed
@Shixiaowei02
Shixiaowei02 deleted the upgrade_nixl_010 branch June 1, 2026 09:52
Shixiaowei02 added a commit to Shixiaowei02/TensorRT-LLM that referenced this pull request Jun 1, 2026
The trtllm dev environment used to install the meta `nixl` package,
which provides a top-level `nixl/__init__.py` shim that redirects to
`nixl_cu13` / `nixl_cu12`. After
NVIDIA#14436 switched the
dependency to `nixl-cu13` directly, the meta shim is no longer present,
so `from nixl import nixl_agent, ...` in
`tensorrt_llm/_torch/disaggregation/nixl/_agent_py.py` raises
`ModuleNotFoundError: No module named 'nixl'` at import time, which in
turn breaks
`test_agent_multi_backends::test_run_with_different_env[1]` during
collection of the spawned subprocess pytest.

Resolve the symbols from `nixl_cu13` (or `nixl_cu12`) first and fall
back to the meta-package only when neither CUDA-specific package is
installed.

Signed-off-by: Shixiaowei02 <39303645+Shixiaowei02@users.noreply.github.com>
Shixiaowei02 added a commit to Shixiaowei02/TensorRT-LLM that referenced this pull request Jun 1, 2026
The trtllm dev environment used to install the meta `nixl` package,
which provides a top-level `nixl/__init__.py` shim that redirects to
`nixl_cu13` / `nixl_cu12`. After
NVIDIA#14436 switched the
dependency to `nixl-cu13` directly, the meta shim is no longer present,
so `from nixl import nixl_agent, ...` in
`tensorrt_llm/_torch/disaggregation/nixl/_agent_py.py` raises
`ModuleNotFoundError: No module named 'nixl'` at import time, which in
turn breaks
`test_agent_multi_backends::test_run_with_different_env[1]` during
collection of the spawned subprocess pytest.

Resolve the symbols from `nixl_cu13` (or `nixl_cu12`) first and fall
back to the meta-package only when neither CUDA-specific package is
installed.

Signed-off-by: Shixiaowei02 <39303645+Shixiaowei02@users.noreply.github.com>
Shixiaowei02 added a commit to Shixiaowei02/TensorRT-LLM that referenced this pull request Jun 2, 2026
The trtllm dev environment used to install the meta `nixl` package,
which provides a top-level `nixl/__init__.py` shim that redirects to
`nixl_cu13` / `nixl_cu12`. After
NVIDIA#14436 switched the
dependency to `nixl-cu13` directly, the meta shim is no longer present,
so `from nixl import nixl_agent, ...` in
`tensorrt_llm/_torch/disaggregation/nixl/_agent_py.py` raises
`ModuleNotFoundError: No module named 'nixl'` at import time, which in
turn breaks
`test_agent_multi_backends::test_run_with_different_env[1]` during
collection of the spawned subprocess pytest.

Resolve the symbols from `nixl_cu13` (or `nixl_cu12`) first and fall
back to the meta-package only when neither CUDA-specific package is
installed.

Signed-off-by: Shixiaowei02 <39303645+Shixiaowei02@users.noreply.github.com>
Shixiaowei02 added a commit to Shixiaowei02/TensorRT-LLM that referenced this pull request Jun 4, 2026
The trtllm dev environment used to install the meta `nixl` package,
which provides a top-level `nixl/__init__.py` shim that redirects to
`nixl_cu13` / `nixl_cu12`. After
NVIDIA#14436 switched the
dependency to `nixl-cu13` directly, the meta shim is no longer present,
so `from nixl import nixl_agent, ...` in
`tensorrt_llm/_torch/disaggregation/nixl/_agent_py.py` raises
`ModuleNotFoundError: No module named 'nixl'` at import time, which in
turn breaks
`test_agent_multi_backends::test_run_with_different_env[1]` during
collection of the spawned subprocess pytest.

Resolve the symbols from `nixl_cu13` (or `nixl_cu12`) first and fall
back to the meta-package only when neither CUDA-specific package is
installed.

Signed-off-by: Shixiaowei02 <39303645+Shixiaowei02@users.noreply.github.com>
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.

5 participants