When users launch with --num-gpus >= 2 without specifying any parallelism flags (--sp-degree, --ulysses-degree, --ring-degree, --enable-cfg-parallel), automatically enable CFG parallel instead of the previous default of pure sequence parallelism (ulysses).

Benchmarked on H200 (2 GPU, warmup excluded):

  Qwen-Image 1024x1024:  SP 11.21s / TP 13.43s / CFG  7.14s (-36%)
  Wan-1.3B T2V 480x832:  SP  5.95s /            CFG  4.77s (-20%)
  Wan-14B T2V 480x832:   SP 23.75s /            CFG 21.27s (-10%)

Users can opt out by explicitly setting --sp-degree or --ulysses-degree.

Motivation

Modifications

Accuracy Tests

Speed Tests and Profiling

Checklist

• Format your code according to the Format code with pre-commit.
• Add unit tests according to the Run and add unit tests.
• Update documentation according to Write documentations.
• Provide accuracy and speed benchmark results according to Test the accuracy and Benchmark the speed.
• Follow the SGLang code style guidance.

Review and Merge Process

1. Ping Merge Oncalls to start the process. See the PR Merge Process.
2. Get approvals from CODEOWNERS and other reviewers.
3. Trigger CI tests with comments or contact authorized users to do so.
   • Common commands include /tag-and-rerun-ci, /tag-run-ci-label, /rerun-failed-ci
4. After green CI and required approvals, ask Merge Oncalls or people with Write permission to merge the PR.