[perf] fix: fix npu profiling scripts

[tongtong0613]

What does this PR do?

fix npu profiling scripts

Checklist Before Starting

• Search for similar PRs. Paste at least one query link here: ...
• Format the PR title as [{modules}] {type}: {description} (This will be checked by the CI)
  • {modules} include fsdp, megatron, veomni, sglang, vllm, rollout, trainer, ci, training_utils, recipe, hardware, deployment, ray, worker, single_controller, misc, perf, model, algo, env, tool, ckpt, doc, data, cfg, reward
  • If this PR involves multiple modules, separate them with , like [megatron, fsdp, doc]
  • {type} is in feat, fix, refactor, chore, test
  • If this PR breaks any API (CLI arguments, config, function signature, etc.), add [BREAKING] to the beginning of the title.
  • Example: [BREAKING][fsdp, megatron] feat: dynamic batching

Test

For changes that can not be tested by CI (e.g., algorithm implementation, new model support), validate by experiment(s) and show results like training curve plots, evaluation results, etc.

API and Usage Example

Demonstrate how the API changes if any, and provide usage example(s) if possible.

# Add code snippet or script demonstrating how to use this

Design & Code Changes

Demonstrate the high-level design if this PR is complex, and list the specific changes.

Checklist Before Submitting

Important

Please check all the following items before requesting a review, otherwise the reviewer might deprioritize this PR for review.

• Read the Contribute Guide.
• Apply pre-commit checks: pre-commit install && pre-commit run --all-files --show-diff-on-failure --color=always
• Add / Update the documentation.
• Add unit or end-to-end test(s) to the CI workflow to cover all the code. If not feasible, explain why: ...
• Once your PR is ready for CI, send a message in the ci-request channel in the verl Slack workspace. (If not accessible, please try the Feishu group (飞书群).)
• If your PR is related to the recipe submodule, please also update the reference to the submodule commit via git submodule update --remote or cd recipe && git pull origin main.

[gemini-code-assist]

Code Review

This pull request updates NPU profiling scripts by increasing gpu_memory_utilization and setting a large update_weights_bucket_megabytes. While these changes can improve performance, setting the weight update bucket size to 4GB in an example script is risky as it significantly increases memory pressure on the training worker and may lead to out-of-memory errors on hardware with less memory. I've added comments suggesting to add a warning about the high memory requirement to mitigate this risk for other users.

[gemini-code-assist]

Setting update_weights_bucket_megabytes to 4096MB allocates a 4GB buffer on the training worker's device. This large allocation, combined with memory for the model and optimizer, significantly increases the risk of out-of-memory (OOM) errors, especially on devices with less memory. While it can improve performance, this high value in an example script is risky. It would be beneficial to add a comment to warn users about the high memory requirement.

Suggested change

	actor_rollout_ref.rollout.checkpoint_engine.update_weights_bucket_megabytes=4096 \
	# NOTE: A large bucket size (4GB) is used for weight updates to improve throughput.
	# This may cause OOM on devices with less memory. Consider lowering if you encounter OOM errors.
	actor_rollout_ref.rollout.checkpoint_engine.update_weights_bucket_megabytes=4096 \

[gemini-code-assist]

Setting update_weights_bucket_megabytes to 4096MB allocates a 4GB buffer on the training worker's device. This large allocation, combined with memory for the model and optimizer, significantly increases the risk of out-of-memory (OOM) errors, especially on devices with less memory. While it can improve performance, this high value in an example script is risky. It would be beneficial to add a comment to warn users about the high memory requirement.

Suggested change

	actor_rollout_ref.rollout.checkpoint_engine.update_weights_bucket_megabytes=4096 \
	# NOTE: A large bucket size (4GB) is used for weight updates to improve throughput.
	# This may cause OOM on devices with less memory. Consider lowering if you encounter OOM errors.
	actor_rollout_ref.rollout.checkpoint_engine.update_weights_bucket_megabytes=4096 \