⚡ WebGPU Benchmark Results (36.11x speedup)
#71
by
Branchverse
- opened
| Batch Size | WASM (fp32) | WebGPU (fp32) |
| 1 | 559.80 | 14.90 |
| 2 | 1104.10 | 66.40 |
| 4 | 2137.40 | 58.50 |
| 8 | 4303.60 | 148.90 |
| 16 | 8648.90 | 241.40 |
| 32 | 17242.40 | 477.50 |
- Model: Xenova/all-MiniLM-L6-v2
- Tests run: WASM (fp32), 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=