-
Notifications
You must be signed in to change notification settings - Fork 660
[PD Disaggregation] Add unittest for splitwise deployment with using rdma #5189
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.
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
base: develop
Are you sure you want to change the base?
Conversation
|
Thanks for your contribution! |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Pull request overview
This pull request adds RDMA (Remote Direct Memory Access) support for splitwise deployment in PD (Prefill-Decode) disaggregation scenarios. The PR introduces a new end-to-end test for RDMA-based cache transfer and includes a shell script utility for detecting RDMA-capable network interfaces.
Key Changes:
- Added RDMA network interface detection script (
get_rdma_nics.sh) to identify appropriate NICs for different hardware configurations - Implemented comprehensive E2E test (
test_ernie_03b_pd_router_v1_rdma.py) for splitwise deployment using RDMA protocol for cache transfer - Fixed minor issues in existing IPC-based test (corrected log file names and function call)
Reviewed changes
Copilot reviewed 3 out of 3 changed files in this pull request and generated 17 comments.
| File | Description |
|---|---|
| tests/e2e/utils/get_rdma_nics.sh | New shell script that detects RDMA-capable network interfaces for GPU, XPU, CPU, and virtual configurations to enable RDMA-based cache transfer |
| tests/e2e/test_ernie_03b_pd_router_v1_rdma.py | New E2E test suite that validates splitwise deployment with RDMA protocol, testing both streaming and non-streaming chat/completion endpoints |
| tests/e2e/test_ernie_03b_pd_router_v1_ipc.py | Minor improvements: updated comment for clarity, fixed log filenames for consistency, and corrected cleanup function call |
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## develop #5189 +/- ##
==========================================
Coverage ? 59.74%
==========================================
Files ? 317
Lines ? 38795
Branches ? 5842
==========================================
Hits ? 23177
Misses ? 13776
Partials ? 1842
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
Motivation
Add unittest
拆分PR: https://github.com/PaddlePaddle/FastDeploy/pull/5027/files
Modifications
Add unittest
Usage or Command
不变
Accuracy Tests
不需要
Checklist
[FDConfig],[APIServer],[Engine],[Scheduler],[PD Disaggregation],[Executor],[Graph Optimization],[Speculative Decoding],[RL],[Models],[Quantization],[Loader],[OP],[KVCache],[DataProcessor],[BugFix],[Docs],[CI],[Optimization],[Feature],[Benchmark],[Others],[XPU],[HPU],[GCU],[DCU],[Iluvatar],[Metax]]pre-commitbefore commit.releasebranch, make sure the PR has been submitted to thedevelopbranch, then cherry-pick it to thereleasebranch with the[Cherry-Pick]PR tag.