Skip to main content
This page reports Ring-SP performance for Wan2.2-TI2V-5B-Diffusers using:
  • Parallel config: sp=2, ulysses=1, ring=2 (short: u1r2)
  • Baseline config: sp=1, ulysses=1, ring=1 (short: u1r1)

​
Benchmark Setup

  • Model: Wan2.2-TI2V-5B-Diffusers
  • GPU: 48G RTX40 series * 2

​
Online Serving

​
Ring SP (u1r2)

sglang serve \
  --model-type diffusion \
  --model-path /model/HuggingFace/Wan-AI/Wan2.2-TI2V-5B-Diffusers \
  --num-gpus 2 --sp-degree 2 --ulysses-degree 1 --ring-degree 2 \
  --port 8898

​
Baseline (u1r1)

sglang serve \
  --model-type diffusion \
  --model-path /model/HuggingFace/Wan-AI/Wan2.2-TI2V-5B-Diffusers \
  --num-gpus 1 --sp-degree 1 --ulysses-degree 1 --ring-degree 1 \
  --port 8898

​
Benchmarks

​
Benchmark Disclaimer

These benchmarks are provided for reference under one specific setup and command configuration. Actual performance may vary with model settings, runtime environment, and request patterns.

​
Stage Time Breakdown

Stage / Metricu1r2 (s)u1r1 baseline (s)Speedup
InputValidation0.10600.10290.97x
TextEncoding1.39652.22611.59x
LatentPreparation0.00020.00021.00x
TimestepPreparation0.00030.00041.33x
Denoising52.635871.67851.36x
Decoding7.670813.43141.75x
Total63.7490.631.42x

​
Memory Usage

Memory Metricu1r2 (GB)u1r1 baseline (GB)Delta
Peak GPU Memory20.0727.40-7.33
Peak Allocated13.3520.40-7.05
Memory Overhead6.727.00-0.28
Overhead Ratio33.5%25.6%+7.9pp

​
Summary

  • End-to-end latency improves from 90.63s to 63.74s (1.42x).
  • Main gains come from Denoising (1.36x) and Decoding (1.75x).
  • Absolute memory usage drops noticeably on Ring-SP (Peak GPU Memory -7.33GB, Peak Allocated -7.05GB).
  • Overhead ratio rises (+7.9pp), so future tuning can focus on reducing communication/runtime overhead while preserving the latency gain.