Skip to content

Revert "[CI] Add Async Eplb nightly CI tests" #30086

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

Closed
SageMoore wants to merge 1 commit into from
Closed

Revert "[CI] Add Async Eplb nightly CI tests" #30086

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
73 changes: 0 additions & 73 deletions .buildkite/scripts/scheduled_integration_test/deepseek_v2_lite_ep_async_eplb.sh
View file
Open in desktop

This file was deleted.

1 change: 0 additions & 1 deletion .buildkite/scripts/scheduled_integration_test/deepseek_v2_lite_ep_eplb.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -50,7 +50,6 @@ for BACK in "${BACKENDS[@]}"; do
--data-parallel-size 2 \
--enable-expert-parallel \
--enable-eplb \
--eplb-config '{"window_size":200,"step_interval":600}' \
--trust-remote-code \
--max-model-len 2048 \
--port $PORT &
Expand Down
74 changes: 0 additions & 74 deletions .buildkite/scripts/scheduled_integration_test/qwen3_next_mtp_async_eplb.sh
View file
Open in desktop

This file was deleted.

20 changes: 1 addition & 19 deletions .buildkite/test-pipeline.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1374,22 +1374,4 @@ steps:
num_gpus: 2
working_dir: "/vllm-workspace"
commands:
- bash .buildkite/scripts/scheduled_integration_test/qwen30b_a3b_fp8_block_ep_eplb.sh 0.8 200 8020 2 1

- label: DeepSeek V2-Lite Async EPLB Accuracy
timeout_in_minutes: 60
gpu: h100
optional: true
num_gpus: 4
working_dir: "/vllm-workspace"
commands:
- bash .buildkite/scripts/scheduled_integration_test/deepseek_v2_lite_ep_async_eplb.sh 0.25 1319 8030

- label: Qwen3-Next-80B-A3B-Instruct MTP Async EPLB Accuracy
timeout_in_minutes: 60
gpu: h100
optional: true
num_gpus: 4
working_dir: "/vllm-workspace"
commands:
- bash .buildkite/scripts/scheduled_integration_test/qwen3_next_mtp_async_eplb.sh 0.8 1319 8040
- bash .buildkite/scripts/scheduled_integration_test/qwen30b_a3b_fp8_block_ep_eplb.sh 0.8 200 8020 2 1
3 changes: 3 additions & 0 deletions vllm/distributed/eplb/rebalance_execute.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -322,6 +322,9 @@ async def transfer_layer(
num_local_physical_experts = next(iter(expert_weights[0])).shape[0]
assert new_global_expert_indices.shape == (num_moe_layers, num_physical_experts)
assert num_physical_experts == ep_size * num_local_physical_experts
# A buffer to hold the expert weights in one layer during the exchange.
# NOTE: Currently we assume the same weights across different layers
# have the same shape.
Comment on lines +325 to +327
Copy link
Copy Markdown
Contributor

@gemini-code-assist gemini-code-assist Bot Dec 4, 2025

Choose a reason for hiding this comment

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

high

This pull request is intended to be a revert, but it includes the addition of these new comments. To maintain a clean and atomic Git history, it's a best practice for a revert to only contain the reverted changes. Please consider moving this documentation improvement to a separate pull request.

Copy link
Copy Markdown
Contributor Author

@SageMoore SageMoore Dec 4, 2025

Choose a reason for hiding this comment

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

This PR is exclusively a reversion of #29385. I have made no other changes.


is_unchanged, is_received_locally, experts_recv_loc = move_to_buffer(
num_local_experts=num_local_physical_experts,
Expand Down