Skip to content

chore:[BREAKING CHANGE] use cacheTransceiverConfig as knobs for disagg service #5234

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Merged
merged 7 commits into from
Jul 17, 2025
Merged

chore:[BREAKING CHANGE] use cacheTransceiverConfig as knobs for disagg service #5234

merged 7 commits into from
Jul 17, 2025

Conversation

chuangz0
Copy link
Collaborator

@chuangz0 chuangz0 commented Jun 16, 2025 •
edited by coderabbitai bot
Loading

PR title

Please write the PR title by following template:

[JIRA ticket link/nvbug link/github issue link][fix/feat/doc/infra/...] <summary of this PR>

For example, assume I have a PR hope to support a new feature about cache manager of Jira TRTLLM-1000 ticket, it would be like

[TRTLLM-1000][feat] Support a new feature about cache manager

Description

Please explain the issue and the solution in short.

We use the cache_transceiver_config configuration to set up disaggregated serving, which includes the following parameters:

cache_transceiver_config:
  backend: <str>
  max_tokens_in_buffer: <int>

Test Coverage

GitHub Bot Help

/bot [-h] ['run', 'kill', 'skip', 'reuse-pipeline'] ...

Provide a user friendly way for developers to interact with a Jenkins server.

Run /bot [-h|--help] to print this help message.

See details below for each supported subcommand.

run [--disable-fail-fast --skip-test --stage-list "A10-1, xxx" --gpu-type "A30, H100_PCIe" --add-multi-gpu-test --only-multi-gpu-test --disable-multi-gpu-test --post-merge --extra-stage "H100_PCIe-[Post-Merge]-1, xxx"]

Launch build/test pipelines. All previously running jobs will be killed.

--disable-fail-fast (OPTIONAL) : Disable fail fast on build/tests/infra failures.

--skip-test (OPTIONAL) : Skip all test stages, but still run build stages, package stages and sanity check stages. Note: Does NOT update GitHub check status.

--stage-list "A10-1, xxx" (OPTIONAL) : Only run the specified test stages. Examples: "A10-1, xxx". Note: Does NOT update GitHub check status.

--gpu-type "A30, H100_PCIe" (OPTIONAL) : Only run the test stages on the specified GPU types. Examples: "A30, H100_PCIe". Note: Does NOT update GitHub check status.

--only-multi-gpu-test (OPTIONAL) : Only run the multi-GPU tests. Note: Does NOT update GitHub check status.

--disable-multi-gpu-test (OPTIONAL) : Disable the multi-GPU tests. Note: Does NOT update GitHub check status.

--add-multi-gpu-test (OPTIONAL) : Force run the multi-GPU tests. Will also run L0 pre-merge pipeline.

--post-merge (OPTIONAL) : Run the L0 post-merge pipeline instead of the ordinary L0 pre-merge pipeline.

--extra-stage "H100_PCIe-[Post-Merge]-1, xxx" (OPTIONAL) : Run the ordinary L0 pre-merge pipeline and specified test stages. Examples: --extra-stage "H100_PCIe-[Post-Merge]-1, xxx".

For guidance on mapping tests to stage names, see docs/source/reference/ci-overview.md.

kill

kill

Kill all running builds associated with pull request.

skip

skip --comment COMMENT

Skip testing for latest commit on pull request. --comment "Reason for skipping build/test" is required. IMPORTANT NOTE: This is dangerous since lack of user care and validation can cause top of tree to break.

reuse-pipeline

reuse-pipeline

Reuse a previous pipeline to validate current commit. This action will also kill all currently running builds associated with the pull request. IMPORTANT NOTE: This is dangerous since lack of user care and validation can cause top of tree to break.

Summary by CodeRabbit

  • New Features

    • Introduced explicit backend selection for cache transceiver configuration, supporting DEFAULT, UCX, NIXL, and MPI backends.
    • Added configuration options for specifying cache transceiver backend and buffer size in both Python and YAML interfaces.

  • Bug Fixes

    • Improved error messages and logging for misconfigurations or deprecated backend usage.

  • Documentation

    • Updated guides and examples to reflect new backend configuration options and removed outdated environment variable instructions.
    • Added new FAQ and clarified known issues regarding backend selection and performance.

  • Tests

    • Updated and extended test cases and integration configs to use the new cache transceiver backend configuration.
    • Added tests for multiple backend types (UCX, MPI, NIXL) and improved serialization checks.

  • Refactor

    • Simplified interface by removing legacy communication type enums and centralizing backend selection in configuration objects.
    • Enhanced buffer size calculation to account for per-window cache sizes and backend-specific settings.

  • Chores

    • Updated configuration files and test lists to align with the new backend configuration approach.

@chuangz0 chuangz0 requested a review from a team as a code owner June 16, 2025 07:07
@chuangz0 chuangz0 requested a review from Naveassaf June 16, 2025 07:07
@chuangz0 chuangz0 force-pushed the cache_transceier_default branch from cfaaffe to 66b9964 Compare June 16, 2025 09:45
@chuangz0
Copy link
Collaborator Author

/bot run --add-multi-gpu-test

@tensorrt-cicd
Copy link
Collaborator

PR_Github #9015 [ run ] triggered by Bot

@tensorrt-cicd
Copy link
Collaborator

PR_Github #9015 [ run ] completed with state FAILURE
/LLM/main/L0_MergeRequest_PR pipeline #6585 completed with status: 'FAILURE'

pcastonguay
pcastonguay requested changes Jun 16, 2025
View reviewed changes
@chuangz0 chuangz0 removed the request for review from Naveassaf June 17, 2025 01:43
@chuangz0 chuangz0 force-pushed the cache_transceier_default branch 2 times, most recently from 24f5716 to 2372f9c Compare June 17, 2025 07:52
@chuangz0
Copy link
Collaborator Author

/bot run --add-multi-gpu-test

@tensorrt-cicd
Copy link
Collaborator

PR_Github #9157 [ run ] triggered by Bot

@pcastonguay pcastonguay changed the title use cacheTransceiverConfig as knobs for diagg service chore: use cacheTransceiverConfig as knobs for disagg service Jun 17, 2025
@chuangz0 chuangz0 force-pushed the cache_transceier_default branch from 5f549dc to 067a5d6 Compare June 18, 2025 01:46
@chuangz0
Copy link
Collaborator Author

/bot run --add-multi-gpu-test

@tensorrt-cicd
Copy link
Collaborator

PR_Github #9271 [ run ] triggered by Bot

Shixiaowei02
Shixiaowei02 reviewed Jun 18, 2025
View reviewed changes
Shixiaowei02
Shixiaowei02 reviewed Jun 18, 2025
View reviewed changes
Shixiaowei02
Shixiaowei02 reviewed Jun 18, 2025
View reviewed changes
@tensorrt-cicd
Copy link
Collaborator

PR_Github #9271 [ run ] completed with state SUCCESS
/LLM/main/L0_MergeRequest_PR pipeline #6800 completed with status: 'SUCCESS'
Pipeline passed with automatic retried tests. Check the rerun report for details.

@chuangz0 chuangz0 changed the title chore: use cacheTransceiverConfig as knobs for disagg service chore:[BREAKING CHANGING] use cacheTransceiverConfig as knobs for disagg service Jun 19, 2025
@chuangz0 chuangz0 changed the title chore:[BREAKING CHANGING] use cacheTransceiverConfig as knobs for disagg service chore:[BREAKING CHANGE] use cacheTransceiverConfig as knobs for disagg service Jun 19, 2025
@chuangz0
Copy link
Collaborator Author

/bot run

@chuangz0 chuangz0 force-pushed the cache_transceier_default branch from b310055 to cd1500d Compare June 19, 2025 03:08
@tensorrt-cicd
Copy link
Collaborator

PR_Github #9436 [ run ] triggered by Bot

@chuangz0 chuangz0 requested a review from qiaoxj07 June 19, 2025 03:16
@tensorrt-cicd
Copy link
Collaborator

PR_Github #12170 Bot args parsing error: usage: /bot [-h]
{run,kill,skip,submit,reviewers,reuse-pipeline,reuse-review} ...
/bot: error: unrecognized arguments: CI has been passed

@chuangz0
Copy link
Collaborator Author

/bot skip --comment "CI has been passed"

@github-actions GitHub Actions
Copy link

GitHub Bot Help

/bot [-h] ['run', 'kill', 'skip', 'reuse-pipeline'] ...

Provide a user friendly way for developers to interact with a Jenkins server.

Run /bot [-h|--help] to print this help message.

See details below for each supported subcommand.

run [--disable-fail-fast --skip-test --stage-list "A10-1, xxx" --gpu-type "A30, H100_PCIe" --add-multi-gpu-test --only-multi-gpu-test --disable-multi-gpu-test --post-merge --extra-stage "H100_PCIe-[Post-Merge]-1, xxx"]

Launch build/test pipelines. All previously running jobs will be killed.

--disable-fail-fast (OPTIONAL) : Disable fail fast on build/tests/infra failures.

--skip-test (OPTIONAL) : Skip all test stages, but still run build stages, package stages and sanity check stages. Note: Does NOT update GitHub check status.

--stage-list "A10-1, xxx" (OPTIONAL) : Only run the specified test stages. Examples: "A10-1, xxx". Note: Does NOT update GitHub check status.

--gpu-type "A30, H100_PCIe" (OPTIONAL) : Only run the test stages on the specified GPU types. Examples: "A30, H100_PCIe". Note: Does NOT update GitHub check status.

--only-multi-gpu-test (OPTIONAL) : Only run the multi-GPU tests. Note: Does NOT update GitHub check status.

--disable-multi-gpu-test (OPTIONAL) : Disable the multi-GPU tests. Note: Does NOT update GitHub check status.

--add-multi-gpu-test (OPTIONAL) : Force run the multi-GPU tests. Will also run L0 pre-merge pipeline.

--post-merge (OPTIONAL) : Run the L0 post-merge pipeline instead of the ordinary L0 pre-merge pipeline.

--extra-stage "H100_PCIe-[Post-Merge]-1, xxx" (OPTIONAL) : Run the ordinary L0 pre-merge pipeline and specified test stages. Examples: --extra-stage "H100_PCIe-[Post-Merge]-1, xxx".

kill

kill

Kill all running builds associated with pull request.

skip

skip --comment COMMENT

Skip testing for latest commit on pull request. --comment "Reason for skipping build/test" is required. IMPORTANT NOTE: This is dangerous since lack of user care and validation can cause top of tree to break.

reuse-pipeline

reuse-pipeline

Reuse a previous pipeline to validate current commit. This action will also kill all currently running builds associated with the pull request. IMPORTANT NOTE: This is dangerous since lack of user care and validation can cause top of tree to break.

@chuangz0
Copy link
Collaborator Author

/bot skip --comment "CI test has pass"

@tensorrt-cicd
Copy link
Collaborator

PR_Github #12175 [ skip ] triggered by Bot

@tensorrt-cicd
Copy link
Collaborator

PR_Github #12175 [ skip ] completed with state SUCCESS
Skipping testing for commit d39e8c2

@chuangz0 chuangz0 merged commit 44c70c8 into NVIDIA:main Jul 17, 2025
3 checks passed
This was referenced Jul 18, 2025
Closed
Merged
Merged
reasonsolo pushed a commit to reasonsolo/TensorRT-LLM that referenced this pull request Jul 21, 2025
@chuangz0 @reasonsolo
chore:[BREAKING CHANGE] use cacheTransceiverConfig as knobs for disag…
e0187ad 
…g service (NVIDIA#5234)

Signed-off-by: Chuang Zhu <[email protected]>
This was referenced Jul 22, 2025
Closed
Merged
Closed
Draft
Merged
@pcastonguay pcastonguay mentioned this pull request Jul 28, 2025
Closed
4 tasks
This was referenced Aug 4, 2025
Closed
Merged
Merged
Open
This was referenced Aug 11, 2025
Open
Open
Merged
Open
Open
Draft
Open
@coderabbitai coderabbitai bot mentioned this pull request Aug 19, 2025
Draft
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@Superjomn Superjomn Superjomn left review comments

@coderabbitai coderabbitai[bot] coderabbitai[bot] left review comments

@Shixiaowei02 Shixiaowei02 Shixiaowei02 approved these changes

@pcastonguay pcastonguay pcastonguay approved these changes

@nv-guomingz nv-guomingz nv-guomingz approved these changes

@qiaoxj07 qiaoxj07 Awaiting requested review from qiaoxj07

@juney-nvidia juney-nvidia Awaiting requested review from juney-nvidia juney-nvidia is a code owner automatically assigned from NVIDIA/trt-llm-llmapi-devs

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
6 participants
@chuangz0 @tensorrt-cicd @Superjomn @Shixiaowei02 @pcastonguay @nv-guomingz