⚡ WebGPU Benchmark Results (84.50x speedup)
#89
by
Xenova
HF staff
- opened
Batch Size | WASM (fp32) | WebGPU (fp32) |
1 | 2164.80 | 221.40 |
2 | 4230.30 | 216.40 |
4 | 8482.10 | 398.00 |
8 | 16882.80 | 467.50 |
16 | 33591.60 | 521.10 |
32 | 66389.60 | 785.70 |
- Model: Snowflake/snowflake-arctic-embed-s
- 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/123.0.0.0 Safari/537.36
- GPU: vendor=nvidia, architecture=turing, device=, description=