xref: /aosp_15_r20/external/pytorch/benchmarks/inference/results/output_128_true.md (revision da0073e96a02ea20f0ac840b70461e3646d07c45) 
1## Batch Size 128 Compile true
2
3| Experiment | Warmup_latency (s) | Average_latency (s) | Throughput (samples/sec) | GPU Utilization (%) |
4| ---------- | ------------------ | ------------------- | ------------------------ | ------------------- |
5| original | 14.358 +/- 0.981 | 14.250 +/- 0.758 | 522.998 +/- 20.830 | 55.501 +/- 2.123 |
6| h2d_d2h_threads | 12.520 +/- 0.253 | 12.774 +/- 0.714 | 600.578 +/- 27.662 | 61.534 +/- 3.653 |
7