List your company

Grouped-Query Attention (GQA) on NVIDIA A100 GPU Synthetic Performance

154.7TFLOPSDeepSeek-V3 + Ours
-1.61238.96979.55120.131Jun 14, 2025

Evaluation results

MethodLinks
DeepSeek-V3 + OursHead Dimension=64, Sequence Length=2k, Causal Mask=true2025.06154.735.16
DeepSeek-V3Head Dimension=64, Sequence Length=2k, Causal Mask=true2025.064.4—