Skip to content

Commit

Permalink
Add benchmark result of T-Head TH1520 (Tencent#4240)
Browse files Browse the repository at this point in the history
`cpuinfo`: 

```
isa             : rv64imafdcvsu
mmu             : sv39
cpu-freq                : 1.848Ghz
cpu-icache              : 64KB
cpu-dcache              : 64KB
cpu-l2cache             : 1MB
cpu-tlb         : 1024 4-ways
cpu-cacheline           : 64Bytes
cpu-vector              : 0.7.1
```

Compiled with `-DCMAKE_TOOLCHAIN_FILE=../toolchains/c910-v240.toolchain.cmake -DCMAKE_BUILD_TYPE=release -DNCNN_OPENMP=OFF -DNCNN_THREADS=OFF -DNCNN_RUNTIME_CPU=OFF -DNCNN_RVV=ON -DNCNN_SIMPLEOCV=ON -DNCNN_BUILD_EXAMPLES=ON` 

Seems much worse than expected 🤔
  • Loading branch information
YuzukiTsuru authored Oct 6, 2022
1 parent 59a6fa3 commit eb9bb5d
Showing 1 changed file with 34 additions and 0 deletions.
34 changes: 34 additions & 0 deletions benchmark/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -4370,3 +4370,37 @@ cooling_down = 1
vision_transformer min = 22201.32 max = 22510.75 avg = 22315.09
FastestDet min = 146.94 max = 148.50 avg = 147.44
```

### T-Head TH1520 (C910V<aka C920>, 1.848 GHz x 4)

Tested on `Linux anolis-riscv 5.10.112-00579-g8e3db308d5a5 #23 SMP PREEMPT Fri Aug 12 10:17:32 CST 2022 riscv64 riscv64 riscv64 GNU/Linux`

```
[root@anolis-riscv benchmark]# ./benchncnn
syscall error -1
loop_count = 4
num_threads = 4
powersave = 0
gpu_device = -1
cooling_down = 1
squeezenet min = 187.88 max = 188.82 avg = 188.13
squeezenet_int8 min = 2388.26 max = 2446.92 avg = 2411.46
mobilenet min = 321.46 max = 323.34 avg = 322.19
mobilenet_int8 min = 2318.93 max = 2458.55 avg = 2400.99
mobilenet_v2 min = 214.01 max = 216.00 avg = 215.35
mobilenet_v3 min = 247.71 max = 248.18 avg = 247.96
shufflenet min = 155.58 max = 155.85 avg = 155.67
shufflenet_v2 min = 99.50 max = 99.75 avg = 99.63
mnasnet min = 261.46 max = 263.83 avg = 262.53
proxylessnasnet min = 315.40 max = 316.89 avg = 316.28
efficientnet_b0 min = 484.97 max = 486.16 avg = 485.55
efficientnetv2_b0 min = 453.03 max = 453.40 avg = 453.21
regnety_400m min = 314.09 max = 315.33 avg = 314.77
blazeface min = 46.14 max = 46.69 avg = 46.39
googlenet min = 650.99 max = 653.60 avg = 651.69
googlenet_int8 min = 5435.11 max = 6391.98 avg = 6012.81
resnet18 min = 505.48 max = 506.70 avg = 506.06
resnet18_int8 min = 5053.33 max = 6599.94 avg = 6001.86
alexnet min = 403.68 max = 404.60 avg = 404.23
vgg16 min = 2731.55 max = 2746.48 avg = 2738.82
```

0 comments on commit eb9bb5d

Please sign in to comment.