⚡ WebGPU Benchmark Results (93.82x speedup) – M1 Max Xenova/gte-base
#63
by
pcuenq
- opened
| Batch Size | WASM (fp32) | WebGPU (fp16) | WebGPU (fp32) |
| 1 | 2594.80 | 70.20 | 64.80 |
| 2 | 5171.00 | 96.20 | 123.10 |
| 4 | 10495.50 | 153.20 | 226.30 |
| 8 | 21334.40 | 273.00 | 434.90 |
| 16 | 43271.70 | 508.20 | 847.70 |
| 32 | 89203.60 | 970.90 | 1671.40 |
| 64 | 178300.00 | 1900.40 | 3324.00 |
- Model: Xenova/gte-base
- Tests run: WASM (fp32), WebGPU (fp16), WebGPU (fp32)
- Sequence length: 512
- Browser: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/122.0.0.0 Safari/537.36
- GPU: vendor=apple, architecture=common-3, device=, description=