Skip to content

perf: Set micro_batch_size=2 for qwen3 235B B300 V2 configs#3252

Merged
ko3n1g merged 1 commit into
mainfrom
rmukundan/qwen3_235_b300_mbs_main
Apr 9, 2026
Merged

perf: Set micro_batch_size=2 for qwen3 235B B300 V2 configs#3252
ko3n1g merged 1 commit into
mainfrom
rmukundan/qwen3_235_b300_mbs_main

Conversation

@rhmukundan

@rhmukundan rhmukundan commented Apr 9, 2026 •
edited by coderabbitai Bot
Loading

Copy link
Copy Markdown
Contributor

Summary by CodeRabbit

  • Chores
    • Updated performance configurations for Qwen3 model training presets to optimize resource allocation and batch processing efficiency.

@rhmukundan
perf: Set micro_batch_size=2 for qwen3 235B B300 V2 configs
9b6e27e 
Signed-off-by: Raghav Hrishikeshan Mukundan <rmukundan@nvidia.com>
@copy-pr-bot

copy-pr-bot Bot commented Apr 9, 2026

Copy link
Copy Markdown

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

@rhmukundan rhmukundan self-assigned this Apr 9, 2026
@rhmukundan rhmukundan added this to the 26.04 milestone Apr 9, 2026
@rhmukundan rhmukundan added the r0.4.0 Auto-cherrypick to release branch. Apply before merge; cherrypick happens after merge. label Apr 9, 2026
@yaoyu-33 yaoyu-33 added feature New capabilities, enhancements, or enablement work area:perf Performance optimizations and benchmarking labels Apr 9, 2026
@dingqingy-nv

dingqingy-nv commented Apr 9, 2026

Copy link
Copy Markdown
Contributor

removing the r0.4.0 label as a manual cherry-pick PR is created at #3251

@dingqingy-nv dingqingy-nv removed the r0.4.0 Auto-cherrypick to release branch. Apply before merge; cherrypick happens after merge. label Apr 9, 2026
@rhmukundan rhmukundan marked this pull request as ready for review April 9, 2026 23:52
@dingqingy-nv dingqingy-nv enabled auto-merge (squash) April 9, 2026 23:54
@ko3n1g ko3n1g disabled auto-merge April 9, 2026 23:54
@ko3n1g ko3n1g merged commit bb767c9 into main Apr 9, 2026
2 checks passed
@ko3n1g ko3n1g deleted the rmukundan/qwen3_235_b300_mbs_main branch April 9, 2026 23:55
@coderabbitai

coderabbitai Bot commented Apr 9, 2026

Copy link
Copy Markdown
Contributor

Caution

Review failed

Pull request was closed or merged during review

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info
⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: c2bd929c-b8f2-41b1-8f07-7ef03aa2b707

📥 Commits

Reviewing files that changed from the base of the PR and between 0dce3d4 and 9b6e27e.

📒 Files selected for processing (1)
  • scripts/performance/configs/qwen/qwen3_workload_base_configs.py
📝 Walkthrough

Walkthrough

This change adds micro_batch_size=2 to two V2 preset configurations for Qwen3 pretraining workloads on B300 hardware, specifically the BF16 and FP8 compressed symmetric variants.

Changes

Cohort / File(s) Summary
Qwen3 Configuration Updates
scripts/performance/configs/qwen/qwen3_workload_base_configs.py		 Added micro_batch_size=2 parameter to QWEN3_235B_A22B_PRETRAIN_CONFIG_B300_BF16_V2 and QWEN3_235B_A22B_PRETRAIN_CONFIG_B300_FP8_CS_V2 configuration presets.

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~3 minutes

🚥 Pre-merge checks | ✅ 4
✅ Passed checks (4 passed)
Check name Status Explanation
Description Check ✅ Passed Check skipped - CodeRabbit’s high-level summary is enabled.
Title check ✅ Passed The title accurately and specifically describes the main change: setting micro_batch_size=2 for qwen3 235B B300 V2 configs, which matches the changeset exactly.
Docstring Coverage ✅ Passed No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Test Results For Major Changes ✅ Passed PR adds minimal configuration parameter adjustment (2 lines) with no code logic modifications, qualifying as minor change under custom check criteria.

✏️ Tip: You can configure your own custom pre-merge checks in the settings.

✨ Finishing Touches
📝 Generate docstrings
  • Create stacked PR
  • Commit on current branch
🧪 Generate unit tests (beta)
  • Create PR with unit tests
  • Commit unit tests in branch rmukundan/qwen3_235_b300_mbs_main

Comment @coderabbitai help to get the list of available commands and usage tips.

yaoyu-33 pushed a commit to conver334/Megatron-Bridge that referenced this pull request Apr 10, 2026
@rhmukundan @yaoyu-33
perf: Set micro_batch_size=2 for qwen3 235B B300 V2 configs (NVIDIA-N…
3b79556 
…eMo#3252)

Signed-off-by: Raghav Hrishikeshan Mukundan <rmukundan@nvidia.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@dingqingy-nv dingqingy-nv dingqingy-nv approved these changes

Assignees

@rhmukundan rhmukundan

Labels

area:perf Performance optimizations and benchmarking feature New capabilities, enhancements, or enablement work

Projects

None yet

Milestone

26.04

Development

Successfully merging this pull request may close these issues.

4 participants

@rhmukundan @dingqingy-nv @ko3n1g @yaoyu-33