Skip to content

Instantly share code, notes, and snippets.

@iree-github-actions-bot

iree-github-actions-bot/iree-full-benchmark-result-1145.md

Created August 26, 2022 02:23
Show Gist options
  • Save iree-github-actions-bot/f1f764b07cb008157429875a396c09c1 to your computer and use it in GitHub Desktop.
Save iree-github-actions-bot/f1f764b07cb008157429875a396c09c1 to your computer and use it in GitHub Desktop.
Download ZIP
Raw
iree-full-benchmark-result-1145.md

Full Benchmark Summary

Improved Latencies 🎉

Benchmark Name Average Latency (ms) Median Latency (ms) Latency Standard Deviation (ms)
PoseNet [fp32] (TFLite) 8-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 6.268 (vs. 8.382, 25.22%↓) 6.277 0.034
PoseNet [fp32] (TFLite) 4-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 7.976 (vs. 10.196, 21.78%↓) 7.970 0.038
PersonDetect [int8] (TFLite) 8-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 1.647 (vs. 1.972, 16.47%↓) 1.647 0.004
DeepLabV3 [fp32] (TFLite) 8-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 9.942 (vs. 11.197, 11.22%↓) 9.947 0.025
DeepLabV3 [fp32] (TFLite) 4-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 12.679 (vs. 14.222, 10.85%↓) 12.670 0.087
PoseNet [fp32] (TFLite) 1-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 20.389 (vs. 21.855, 6.71%↓) 20.399 0.069
PersonDetect [int8] (TFLite) 4-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 1.644 (vs. 1.742, 5.62%↓) 1.646 0.005

Similar Latencies

Benchmark Name Average Latency (ms) Median Latency (ms) Latency Standard Deviation (ms)
DeepLabV3 [fp32] (TFLite) 1-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 30.070 (vs. 31.232, 3.72%↓) 30.029 0.130
MobileNetV3Small [fp32,imagenet] (TFLite) 8-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 3.366 (vs. 3.483, 3.37%↓) 3.366 0.010
MobileNetV3Small [fp32,imagenet] (TFLite) 1-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 4.565 (vs. 4.648, 1.78%↓) 4.562 0.026
MobileNetV2 [fp32,imagenet] (TFLite) 8-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 5.924 (vs. 6.021, 1.62%↓) 5.931 0.033
MobileBertSquad [fp32] (TFLite) full-inference,default-flags with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 185.315 (vs. 188.217, 1.54%↓) 184.707 1.956
MiniLML12H384Uncased [int32] (TF) full-inference,default-flags with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 377.589 (vs. 383.210, 1.47%↓) 378.054 4.878
MiniLML12H384Uncased [int32] (TF) 8-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 64.529 (vs. 65.360, 1.27%↓) 64.577 0.712
EfficientNet [int8] (TFLite) 8-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 24.956 (vs. 25.268, 1.23%↓) 24.963 0.050
MiniLML12H384Uncased [int32] (TF) 4-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 109.560 (vs. 108.310, 1.15%↑) 110.647 1.606
EfficientNet [int8] (TFLite) 1-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 126.140 (vs. 127.505, 1.07%↓) 126.160 0.245
MobileNetV2 [fp32,imagenet] (TFLite) 1-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 13.831 (vs. 13.971, 1.01%↓) 13.799 0.135
MobileSSD [fp32] (TFLite) full-inference,default-flags with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 40.723 (vs. 41.128, 0.98%↓) 40.754 0.323
PoseNet [fp32] (TFLite) full-inference,default-flags with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 20.218 (vs. 20.416, 0.97%↓) 20.237 0.182
DeepLabV3 [fp32] (TFLite) full-inference,default-flags with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 28.830 (vs. 29.102, 0.94%↓) 28.805 0.141
MobileSSD [fp32] (TFLite) 4-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 16.424 (vs. 16.272, 0.93%↑) 16.425 0.055
MobileBertSquad [int8] (TFLite) 8-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 97.623 (vs. 98.527, 0.92%↓) 97.523 0.254
MobileBertSquad [int8] (TFLite) 4-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 154.677 (vs. 153.287, 0.91%↑) 154.624 0.263
MiniLML12H384Uncased [int32] (TF) 1-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 373.842 (vs. 370.489, 0.91%↑) 375.252 3.300
MobileBertSquad [fp32] (TFLite) 4-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 74.455 (vs. 73.798, 0.89%↑) 74.469 0.272
EfficientNet [int8] (TFLite) full-inference,default-flags with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 124.787 (vs. 125.798, 0.80%↓) 124.775 0.084
MobileNetV3Small [fp32,imagenet] (TFLite) full-inference,default-flags with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 4.449 (vs. 4.484, 0.79%↓) 4.422 0.057
MobileNetV3Small [fp32,imagenet] (TFLite) 4-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 3.285 (vs. 3.305, 0.62%↓) 3.280 0.024
EfficientNet [int8] (TFLite) 4-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 40.305 (vs. 40.535, 0.57%↓) 40.319 0.036
MobileNetV2 [fp32,imagenet] (TFLite) full-inference,default-flags with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 13.448 (vs. 13.511, 0.47%↓) 13.477 0.134
MobileBertSquad [fp32] (TFLite) 8-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 53.525 (vs. 53.769, 0.45%↓) 53.547 0.146
MobileSSD [fp32] (TFLite) 1-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 40.973 (vs. 40.872, 0.25%↑) 41.059 0.374
PersonDetect [int8] (TFLite) full-inference,default-flags with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 2.458 (vs. 2.457, 0.07%↑) 2.458 0.003
PersonDetect [int8] (TFLite) 1-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 2.546 (vs. 2.545, 0.06%↑) 2.546 0.002
MobileSSD [fp32] (TFLite) 8-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 12.658 (vs. 12.651, 0.06%↑) 12.666 0.029
MobileBertSquad [int8] (TFLite) full-inference,default-flags with IREE-LLVM-CPU-Sync @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 484.800 (vs. 484.598, 0.04%↑) 484.857 0.933
MobileBertSquad [fp32] (TFLite) 1-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 186.433 (vs. 186.506, 0.04%↓) 186.414 0.597
MobileNetV2 [fp32,imagenet] (TFLite) 4-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 6.860 (vs. 6.858, 0.04%↑) 6.876 0.065
MobileBertSquad [int8] (TFLite) 1-thread,full-inference,default-flags with IREE-LLVM-CPU @ GCP-c2-standard-16 (CPU-x86\_64-CascadeLake) 485.139 (vs. 484.981, 0.03%↑) 485.153 0.988

Regressed Compilation Times 🚩

Benchmark Name Compilation Time (ms)
MobileBertSquad [int8] (TFLite) CPU-RV32-Generic full-inference,default-flags 10999718 (vs. 1369401, 703.25%↑)
MobileBertSquad [int8] (TFLite) CPU-RV64-Generic full-inference,default-flags 395836 (vs. 146289, 170.58%↑)
EfficientNet [int8] (TFLite) CPU-RV32-Generic full-inference,default-flags 17652 (vs. 15420, 14.47%↑)
DeepLabV3 [fp32] (TFLite) CPU-RV64-Generic full-inference,default-flags 7194 (vs. 6459, 11.38%↑)
MobileSSD [fp32] (TFLite) CPU-x86\_64-CascadeLake 8-thread,full-inference,default-flags 25943 (vs. 23462, 10.57%↑)
MobileSSD [fp32] (TFLite) CPU-x86\_64-CascadeLake full-inference,default-flags 25943 (vs. 23462, 10.57%↑)
MobileSSD [fp32] (TFLite) CPU-x86\_64-CascadeLake 4-thread,full-inference,default-flags 25943 (vs. 23462, 10.57%↑)
MobileSSD [fp32] (TFLite) CPU-x86\_64-CascadeLake 1-thread,full-inference,default-flags 25943 (vs. 23462, 10.57%↑)
PersonDetect [int8] (TFLite) CPU-RV64-Generic full-inference,default-flags 7231 (vs. 6589, 9.74%↑)
MobileNetV1 [fp32,imagenet] (TFLite) CPU-RV64-Generic full-inference,default-flags 13900 (vs. 13031, 6.67%↑)
MobileBertSquad [fp32] (TFLite) CPU-x86\_64-CascadeLake 8-thread,full-inference,default-flags 64184 (vs. 60448, 6.18%↑)
MobileBertSquad [fp32] (TFLite) CPU-x86\_64-CascadeLake full-inference,default-flags 64184 (vs. 60448, 6.18%↑)
MobileBertSquad [fp32] (TFLite) CPU-x86\_64-CascadeLake 4-thread,full-inference,default-flags 64184 (vs. 60448, 6.18%↑)
MobileBertSquad [fp32] (TFLite) CPU-x86\_64-CascadeLake 1-thread,full-inference,default-flags 64184 (vs. 60448, 6.18%↑)

Improved Compilation Times 🎉

Benchmark Name Compilation Time (ms)
PoseNet [fp32] (TFLite) CPU-x86\_64-CascadeLake 8-thread,full-inference,default-flags 8767 (vs. 10332, 15.15%↓)
PoseNet [fp32] (TFLite) CPU-x86\_64-CascadeLake full-inference,default-flags 8767 (vs. 10332, 15.15%↓)
PoseNet [fp32] (TFLite) CPU-x86\_64-CascadeLake 4-thread,full-inference,default-flags 8767 (vs. 10332, 15.15%↓)
PoseNet [fp32] (TFLite) CPU-x86\_64-CascadeLake 1-thread,full-inference,default-flags 8767 (vs. 10332, 15.15%↓)
MiniLML12H384Uncased [int32] (TF) CPU-x86\_64-CascadeLake 8-thread,full-inference,default-flags 7391 (vs. 8326, 11.23%↓)
MiniLML12H384Uncased [int32] (TF) CPU-x86\_64-CascadeLake full-inference,default-flags 7391 (vs. 8326, 11.23%↓)
MiniLML12H384Uncased [int32] (TF) CPU-x86\_64-CascadeLake 4-thread,full-inference,default-flags 7391 (vs. 8326, 11.23%↓)
MiniLML12H384Uncased [int32] (TF) CPU-x86\_64-CascadeLake 1-thread,full-inference,default-flags 7391 (vs. 8326, 11.23%↓)
DeepLabV3 [fp32] (TFLite) CPU-x86\_64-CascadeLake 8-thread,full-inference,default-flags 12122 (vs. 13178, 8.01%↓)
DeepLabV3 [fp32] (TFLite) CPU-x86\_64-CascadeLake full-inference,default-flags 12122 (vs. 13178, 8.01%↓)
DeepLabV3 [fp32] (TFLite) CPU-x86\_64-CascadeLake 4-thread,full-inference,default-flags 12122 (vs. 13178, 8.01%↓)
DeepLabV3 [fp32] (TFLite) CPU-x86\_64-CascadeLake 1-thread,full-inference,default-flags 12122 (vs. 13178, 8.01%↓)

Regressed Total Dispatch Sizes 🚩

Benchmark Name Total Dispatch Size (bytes)
MobileBertSquad [int8] (TFLite) CPU-RV32-Generic full-inference,default-flags 84601208 (vs. 22515640, 275.74%↑)
MobileBertSquad [int8] (TFLite) CPU-RV64-Generic full-inference,default-flags 11305392 (vs. 4133184, 173.53%↑)
EfficientNet [int8] (TFLite) CPU-RV32-Generic full-inference,default-flags 749980 (vs. 610652, 22.82%↑)
EfficientNet [int8] (TFLite) CPU-RV64-Generic full-inference,default-flags 182816 (vs. 157696, 15.93%↑)
MobileNetV1 [fp32,imagenet] (TFLite) CPU-RV64-Generic full-inference,default-flags 54032 (vs. 49520, 9.11%↑)
PersonDetect [int8] (TFLite) CPU-x86\_64-CascadeLake 8-thread,full-inference,default-flags 107368 (vs. 102136, 5.12%↑)
PersonDetect [int8] (TFLite) CPU-x86\_64-CascadeLake full-inference,default-flags 107368 (vs. 102136, 5.12%↑)
PersonDetect [int8] (TFLite) CPU-x86\_64-CascadeLake 4-thread,full-inference,default-flags 107368 (vs. 102136, 5.12%↑)
PersonDetect [int8] (TFLite) CPU-x86\_64-CascadeLake 1-thread,full-inference,default-flags 107368 (vs. 102136, 5.12%↑)

All Compilation Metrics

Benchmark Name Compilation Time (ms) Total Dispatch Size (bytes)
MobileBertSquad [fp32] (TFLite) CPU-RV64-Generic full-inference,default-flags 65748 (vs. 62856, 4.60%↑) 38072 (vs. 38008, 0.17%↑)
MobileNetV1 [fp32,imagenet] (TFLite) CPU-RV64-Generic full-inference,default-flags 13900 (vs. 13031, 6.67%↑) 54032 (vs. 49520, 9.11%↑)
MobileBertSquad [int8] (TFLite) CPU-RV32-Generic full-inference,default-flags 10999718 (vs. 1369401, 703.25%↑) 84601208 (vs. 22515640, 275.74%↑)
MobileBertSquad [int8] (TFLite) CPU-RV64-Generic full-inference,default-flags 395836 (vs. 146289, 170.58%↑) 11305392 (vs. 4133184, 173.53%↑)
EfficientNet [int8] (TFLite) CPU-RV32-Generic full-inference,default-flags 17652 (vs. 15420, 14.47%↑) 749980 (vs. 610652, 22.82%↑)
EfficientNet [int8] (TFLite) CPU-RV64-Generic full-inference,default-flags 12037 (vs. 11483, 4.82%↑) 182816 (vs. 157696, 15.93%↑)
DeepLabV3 [fp32] (TFLite) CPU-RV64-Generic full-inference,default-flags 7194 (vs. 6459, 11.38%↑) 45152 (vs. 44816, 0.75%↑)
PersonDetect [int8] (TFLite) CPU-RV32-Generic full-inference,default-flags 8598 (vs. 8379, 2.61%↑) 272712 (vs. 284936, 4.29%↓)
PersonDetect [int8] (TFLite) CPU-RV64-Generic full-inference,default-flags 7231 (vs. 6589, 9.74%↑) 85736 (vs. 87384, 1.89%↓)
MiniLML12H384Uncased [int32] (TF) CPU-x86_64-CascadeLake 8-thread,full-inference,default-flags 7391 (vs. 8326, 11.23%↓) 66376 (vs. 65928, 0.68%↑)
MiniLML12H384Uncased [int32] (TF) CPU-x86_64-CascadeLake full-inference,default-flags 7391 (vs. 8326, 11.23%↓) 66376 (vs. 65928, 0.68%↑)
MiniLML12H384Uncased [int32] (TF) CPU-x86_64-CascadeLake 4-thread,full-inference,default-flags 7391 (vs. 8326, 11.23%↓) 66376 (vs. 65928, 0.68%↑)
MiniLML12H384Uncased [int32] (TF) CPU-x86_64-CascadeLake 1-thread,full-inference,default-flags 7391 (vs. 8326, 11.23%↓) 66376 (vs. 65928, 0.68%↑)
MobileNetV3Small [fp32,imagenet] (TFLite) CPU-x86_64-CascadeLake 8-thread,full-inference,default-flags 17664 (vs. 17604, 0.34%↑) 215432 (vs. 216328, 0.41%↓)
MobileNetV3Small [fp32,imagenet] (TFLite) CPU-x86_64-CascadeLake full-inference,default-flags 17664 (vs. 17604, 0.34%↑) 215432 (vs. 216328, 0.41%↓)
MobileNetV3Small [fp32,imagenet] (TFLite) CPU-x86_64-CascadeLake 4-thread,full-inference,default-flags 17664 (vs. 17604, 0.34%↑) 215432 (vs. 216328, 0.41%↓)
MobileNetV3Small [fp32,imagenet] (TFLite) CPU-x86_64-CascadeLake 1-thread,full-inference,default-flags 17664 (vs. 17604, 0.34%↑) 215432 (vs. 216328, 0.41%↓)
MobileBertSquad [fp32] (TFLite) CPU-x86_64-CascadeLake 8-thread,full-inference,default-flags 64184 (vs. 60448, 6.18%↑) 81864 (vs. 82536, 0.81%↓)
MobileBertSquad [fp32] (TFLite) CPU-x86_64-CascadeLake full-inference,default-flags 64184 (vs. 60448, 6.18%↑) 81864 (vs. 82536, 0.81%↓)
MobileBertSquad [fp32] (TFLite) CPU-x86_64-CascadeLake 4-thread,full-inference,default-flags 64184 (vs. 60448, 6.18%↑) 81864 (vs. 82536, 0.81%↓)
MobileBertSquad [fp32] (TFLite) CPU-x86_64-CascadeLake 1-thread,full-inference,default-flags 64184 (vs. 60448, 6.18%↑) 81864 (vs. 82536, 0.81%↓)
PoseNet [fp32] (TFLite) CPU-x86_64-CascadeLake 8-thread,full-inference,default-flags 8767 (vs. 10332, 15.15%↓) 85064 (vs. 84168, 1.06%↑)
PoseNet [fp32] (TFLite) CPU-x86_64-CascadeLake full-inference,default-flags 8767 (vs. 10332, 15.15%↓) 85064 (vs. 84168, 1.06%↑)
PoseNet [fp32] (TFLite) CPU-x86_64-CascadeLake 4-thread,full-inference,default-flags 8767 (vs. 10332, 15.15%↓) 85064 (vs. 84168, 1.06%↑)
PoseNet [fp32] (TFLite) CPU-x86_64-CascadeLake 1-thread,full-inference,default-flags 8767 (vs. 10332, 15.15%↓) 85064 (vs. 84168, 1.06%↑)
MobileNetV2 [fp32,imagenet] (TFLite) CPU-x86_64-CascadeLake 8-thread,full-inference,default-flags 18086 (vs. 17445, 3.67%↑) 150264 (vs. 150344, 0.05%↓)
MobileNetV2 [fp32,imagenet] (TFLite) CPU-x86_64-CascadeLake full-inference,default-flags 18086 (vs. 17445, 3.67%↑) 150264 (vs. 150344, 0.05%↓)
MobileNetV2 [fp32,imagenet] (TFLite) CPU-x86_64-CascadeLake 4-thread,full-inference,default-flags 18086 (vs. 17445, 3.67%↑) 150264 (vs. 150344, 0.05%↓)
MobileNetV2 [fp32,imagenet] (TFLite) CPU-x86_64-CascadeLake 1-thread,full-inference,default-flags 18086 (vs. 17445, 3.67%↑) 150264 (vs. 150344, 0.05%↓)
MobileBertSquad [int8] (TFLite) CPU-x86_64-CascadeLake 8-thread,full-inference,default-flags 211607 (vs. 204690, 3.38%↑) 7496664 (vs. 7339592, 2.14%↑)
MobileBertSquad [int8] (TFLite) CPU-x86_64-CascadeLake full-inference,default-flags 211607 (vs. 204690, 3.38%↑) 7496664 (vs. 7339592, 2.14%↑)
MobileBertSquad [int8] (TFLite) CPU-x86_64-CascadeLake 4-thread,full-inference,default-flags 211607 (vs. 204690, 3.38%↑) 7496664 (vs. 7339592, 2.14%↑)
MobileBertSquad [int8] (TFLite) CPU-x86_64-CascadeLake 1-thread,full-inference,default-flags 211607 (vs. 204690, 3.38%↑) 7496664 (vs. 7339592, 2.14%↑)
MobileSSD [fp32] (TFLite) CPU-x86_64-CascadeLake 8-thread,full-inference,default-flags 25943 (vs. 23462, 10.57%↑) 264696 (vs. 262776, 0.73%↑)
MobileSSD [fp32] (TFLite) CPU-x86_64-CascadeLake full-inference,default-flags 25943 (vs. 23462, 10.57%↑) 264696 (vs. 262776, 0.73%↑)
MobileSSD [fp32] (TFLite) CPU-x86_64-CascadeLake 4-thread,full-inference,default-flags 25943 (vs. 23462, 10.57%↑) 264696 (vs. 262776, 0.73%↑)
MobileSSD [fp32] (TFLite) CPU-x86_64-CascadeLake 1-thread,full-inference,default-flags 25943 (vs. 23462, 10.57%↑) 264696 (vs. 262776, 0.73%↑)
EfficientNet [int8] (TFLite) CPU-x86_64-CascadeLake 8-thread,full-inference,default-flags 15549 (vs. 15251, 1.95%↑) 214632 (vs. 218328, 1.69%↓)
EfficientNet [int8] (TFLite) CPU-x86_64-CascadeLake full-inference,default-flags 15549 (vs. 15251, 1.95%↑) 214632 (vs. 218328, 1.69%↓)
EfficientNet [int8] (TFLite) CPU-x86_64-CascadeLake 4-thread,full-inference,default-flags 15549 (vs. 15251, 1.95%↑) 214632 (vs. 218328, 1.69%↓)
EfficientNet [int8] (TFLite) CPU-x86_64-CascadeLake 1-thread,full-inference,default-flags 15549 (vs. 15251, 1.95%↑) 214632 (vs. 218328, 1.69%↓)
DeepLabV3 [fp32] (TFLite) CPU-x86_64-CascadeLake 8-thread,full-inference,default-flags 12122 (vs. 13178, 8.01%↓) 151880 (vs. 152120, 0.16%↓)
DeepLabV3 [fp32] (TFLite) CPU-x86_64-CascadeLake full-inference,default-flags 12122 (vs. 13178, 8.01%↓) 151880 (vs. 152120, 0.16%↓)
DeepLabV3 [fp32] (TFLite) CPU-x86_64-CascadeLake 4-thread,full-inference,default-flags 12122 (vs. 13178, 8.01%↓) 151880 (vs. 152120, 0.16%↓)
DeepLabV3 [fp32] (TFLite) CPU-x86_64-CascadeLake 1-thread,full-inference,default-flags 12122 (vs. 13178, 8.01%↓) 151880 (vs. 152120, 0.16%↓)
PersonDetect [int8] (TFLite) CPU-x86_64-CascadeLake 8-thread,full-inference,default-flags 7764 (vs. 7611, 2.01%↑) 107368 (vs. 102136, 5.12%↑)
PersonDetect [int8] (TFLite) CPU-x86_64-CascadeLake full-inference,default-flags 7764 (vs. 7611, 2.01%↑) 107368 (vs. 102136, 5.12%↑)
PersonDetect [int8] (TFLite) CPU-x86_64-CascadeLake 4-thread,full-inference,default-flags 7764 (vs. 7611, 2.01%↑) 107368 (vs. 102136, 5.12%↑)
PersonDetect [int8] (TFLite) CPU-x86_64-CascadeLake 1-thread,full-inference,default-flags 7764 (vs. 7611, 2.01%↑) 107368 (vs. 102136, 5.12%↑)
MiniLML12H384Uncased [int32] (TF) GPU-CUDA-SM_80 full-inference,default-flags 4123 (vs. 4037, 2.13%↑) 151888 (vs. 151888, 0.00%)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment