TVM Benchmark
Jack Yu edited this page Jan 29, 2019
·
10 revisions
Below we tested MobileFaceNet(dim128) on different platforms.
| @TVM | @NCNN | @MXNet | |
|---|---|---|---|
| Macbook r2015 2.2G@ 4 Threads | 6.4ms | No SIMD optimization on x86 | 28ms |
| Firefly Rk 3399@A72 * 1 | 69.24ms | 72ms | [x] |
| Firefly Rk 3399@Mali GPU[fp32] | 70.32ms | [x] | [x] |
| Firefly Rk 3399@Mali GPU[fp16] | 45.20ms | [x] | [x] |
| Raspberry Pi 3b@A53 * 1 | 308.24ms | 163.41ms | [x] |
Below we tested LResNet100E(dim512) on different platforms.
| @TVM | @MXNet(pip install) | OpenCV-DNN | |
|---|---|---|---|
| E5-2650 v4 2.20GHz @ 24 Threads | 45ms | 600ms | TODO |
| Macbook r2015 2.2G@4 threads | 178ms | 433ms | TODO |
| Nvidia GTX 1080 | 8.40ms | 31ms | TODO |
| Nvidia Tesla P40 | 7.12ms | 15.61ms | TODO |