List your company

Sparse Decode Performance on 128K Context irregular sparse decode

76.77Speedup Factorirregular sparse decode
-2.73817.903538.54559.1865May 22, 2026

Evaluation results

MethodLinks
irregular sparse decodeBatch size (B)=16, Sparsity Level (S)=500x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.0576.77—
irregular sparse decodeBatch size (B)=8, Sparsity Level (S)=500x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.0576.14—
irregular sparse decodeBatch size (B)=4, Sparsity Level (S)=500x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.0542.04—
irregular sparse decodeBatch size (B)=16, Sparsity Level (S)=200x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.0537.32—
irregular sparse decodeBatch size (B)=8, Sparsity Level (S)=200x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.0529.64—
irregular sparse decodeBatch size (B)=4, Sparsity Level (S)=200x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.0524.25—
irregular sparse decodeBatch size (B)=16, Sparsity Level (S)=100x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.0520.09—
irregular sparse decodeBatch size (B)=8, Sparsity Level (S)=100x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.0516.82—
irregular sparse decodeBatch size (B)=4, Sparsity Level (S)=100x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.0513.36—
irregular sparse decodeBatch size (B)=1, Sparsity Level (S)=500x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.0511.14—
irregular sparse decodeBatch size (B)=1, Sparsity Level (S)=200x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.0511.05—
irregular sparse decodeBatch size (B)=16, Sparsity Level (S)=50x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.0510.54—
irregular sparse decodeBatch size (B)=1, Sparsity Level (S)=100x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.0510.25—
irregular sparse decodeBatch size (B)=8, Sparsity Level (S)=50x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.058.88—
irregular sparse decodeBatch size (B)=4, Sparsity Level (S)=50x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.057.45—
irregular sparse decodeBatch size (B)=1, Sparsity Level (S)=50x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.055.57—
irregular sparse decodeBatch size (B)=16, Sparsity Level (S)=20x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.054.35—
irregular sparse decodeBatch size (B)=8, Sparsity Level (S)=20x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.053.75—
irregular sparse decodeBatch size (B)=4, Sparsity Level (S)=20x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.053.18—
irregular sparse decodeBatch size (B)=1, Sparsity Level (S)=20x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.052.58—
irregular sparse decodeBatch size (B)=16, Sparsity Level (S)=10x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.052.21—
irregular sparse decodeBatch size (B)=8, Sparsity Level (S)=10x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.051.9—
irregular sparse decodeBatch size (B)=4, Sparsity Level (S)=10x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.051.64—
irregular sparse decodeBatch size (B)=1, Sparsity Level (S)=10x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.051.45—
irregular sparse decodeBatch size (B)=16, Sparsity Level (S)=4x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.050.89—
irregular sparse decodeBatch size (B)=8, Sparsity Level (S)=4x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.050.77—
irregular sparse decodeBatch size (B)=4, Sparsity Level (S)=4x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.050.66—
irregular sparse decodeBatch size (B)=1, Sparsity Level (S)=4x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.050.63—
irregular sparse decodeBatch size (B)=16, Sparsity Level (S)=2x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.050.45—
irregular sparse decodeBatch size (B)=8, Sparsity Level (S)=2x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.050.38—
irregular sparse decodeBatch size (B)=4, Sparsity Level (S)=2x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.050.33—
irregular sparse decodeBatch size (B)=1, Sparsity Level (S)=2x, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.050.32—
FlashInferBatch size (B)=1, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.05—0.19
FlashInferBatch size (B)=4, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.05—0.72
FlashInferBatch size (B)=8, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.05—1.5
FlashInferBatch size (B)=16, Hardware=H100 80GB HBM3, Precision=FP16, Attention Configuration=GQA Hq=32, Hkv=8, D=128, Context Length=128K, Page size=162026.05—3.08