⚡ WebGPU Benchmark Results (2.05x speedup) - Windows fp16 vs. fp32
#60
by
Xenova
HF staff
- opened
| Batch Size | WebGPU (fp16) | WebGPU (fp32) |
| 1 | 13.10 | 14.10 |
| 2 | 49.80 | 59.00 |
| 4 | 34.30 | 56.90 |
| 8 | 107.20 | 146.40 |
| 16 | 158.60 | 227.70 |
| 32 | 348.40 | 714.10 |
- Model: Xenova/all-MiniLM-L6-v2
- Tests run: WebGPU (fp16), WebGPU (fp32)
- Sequence length: 512
- Browser: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/122.0.0.0 Safari/537.36
- GPU: vendor=nvidia, architecture=turing, device=, description=