Attention Operator Performance on MLA 128 Head Dimension Sequence Length 2k
108.2TFLOPSLLM-TL
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| LLM-TLHead Dimension=128, Mask=Causal, GPU=A100, Base Model=DeepSeek-V32025.06 | 108.2 | 8.01 | |
| cuDNNHead Dimension=128, Mask=Causal, GPU=A1002025.06 | 61.1 | — | |
| torchHead Dimension=128, Mask=Causal, GPU=A1002025.06 | 21.7 | — | |
| DeepSeek-V3Head Dimension=128, Mask=Causal, GPU=A1002025.06 | 13.5 | — |