Skip to content

Instantly share code, notes, and snippets.

@iree-github-actions-bot

iree-github-actions-bot/iree-full-benchmark-result-137.md

Created June 7, 2021 15:04
Show Gist options
  • Save iree-github-actions-bot/53cdcbb5a2d4dd2e2ca73d02aac4eb1b to your computer and use it in GitHub Desktop.
Save iree-github-actions-bot/53cdcbb5a2d4dd2e2ca73d02aac4eb1b to your computer and use it in GitHub Desktop.
Download ZIP
Raw
iree-full-benchmark-result-137.md

Full Benchmark Summary

Improved Benchmarks 🎉

Benchmark Name Average Latency (ms) Median Latency (ms) Latency Standard Deviation (ms)
MobileNetV3Small [fp32,imagenet] (TensorFlow) kernel-execution with IREE-Vulkan @ SM-G980F (GPU-Mali-G77) 23 (vs. 28, 17.86%↓) 23 0
MobileNetV2 [fp32,imagenet] (TensorFlow) full-inference with IREE-Vulkan @ SM-G980F (GPU-Mali-G77) 58 (vs. 66, 12.12%↓) 59 12
MobileNetV3Small [fp32,imagenet] (TensorFlow) 3-thread,little-core,full-inference with IREE-Dylib @ Pixel-4 (CPU-ARMv8.2-A) 429 (vs. 458, 6.33%↓) 424 40
MobileNetV2 [fp32,imagenet] (TensorFlow) kernel-execution with IREE-Vulkan @ SM-G980F (GPU-Mali-G77) 31 (vs. 33, 6.06%↓) 31 2

Similar Benchmarks

Benchmark Name Average Latency (ms) Median Latency (ms) Latency Standard Deviation (ms)
MobileNetV3Small [fp32,imagenet] (TensorFlow) 3-thread,little-core,full-inference with IREE-Dylib @ SM-G980F (CPU-ARMv8.2-A) 303 (vs. 316, 4.11%↓) 314 18
MobileNetV3Small [fp32,imagenet] (TensorFlow) full-inference with IREE-Vulkan @ SM-G980F (GPU-Mali-G77) 79 (vs. 76, 3.95%↑) 78 3
MobileNetV3Small [fp32,imagenet] (TensorFlow) 1-thread,little-core,full-inference with IREE-Dylib @ SM-G980F (CPU-ARMv8.2-A) 360 (vs. 373, 3.49%↓) 372 26
MobileNetV3Small [fp32,imagenet] (TensorFlow) 3-thread,big-core,full-inference with IREE-Dylib @ Pixel-4 (CPU-ARMv8.2-A) 122 (vs. 125, 2.40%↓) 123 12
MobileNetV3Small [fp32,imagenet] (TensorFlow) little-core,full-inference with IREE-Dylib-Sync @ Pixel-4 (CPU-ARMv8.2-A) 384 (vs. 393, 2.29%↓) 381 8
MobileNetV2 [fp32,imagenet] (TensorFlow) 3-thread,big-core,full-inference with IREE-Dylib @ Pixel-4 (CPU-ARMv8.2-A) 226 (vs. 231, 2.16%↓) 221 14
MobileNetV2 [fp32,imagenet] (TensorFlow) 1-thread,big-core,full-inference with IREE-Dylib @ SM-G980F (CPU-ARMv8.2-A) 140 (vs. 143, 2.10%↓) 140 3
MobileNetV2 [fp32,imagenet] (TensorFlow) 3-thread,big-core,full-inference with IREE-Dylib @ SM-G980F (CPU-ARMv8.2-A) 191 (vs. 195, 2.05%↓) 193 5
MobileNetV2 [fp32,imagenet] (TensorFlow) 3-thread,little-core,full-inference with IREE-Dylib @ Pixel-4 (CPU-ARMv8.2-A) 998 (vs. 1018, 1.96%↓) 979 62
MobileNetV2 [fp32,imagenet] (TensorFlow) big-core,full-inference with IREE-Dylib-Sync @ SM-G980F (CPU-ARMv8.2-A) 139 (vs. 141, 1.42%↓) 139 1
MobileNetV2 [fp32,imagenet] (TensorFlow) kernel-execution with IREE-Vulkan @ Pixel-4 (GPU-Adreno-640) 105 (vs. 104, 0.96%↑) 104 1
MobileNetV2 [fp32,imagenet] (TensorFlow) 3-thread,little-core,full-inference with IREE-Dylib @ SM-G980F (CPU-ARMv8.2-A) 873 (vs. 866, 0.81%↑) 874 8
MobileNetV3Small [fp32,imagenet] (TensorFlow) 1-thread,little-core,full-inference with IREE-Dylib @ Pixel-4 (CPU-ARMv8.2-A) 390 (vs. 392, 0.51%↓) 391 2
MobileNetV2 [fp32,imagenet] (TensorFlow) big-core,full-inference with IREE-Dylib-Sync @ Pixel-4 (CPU-ARMv8.2-A) 211 (vs. 210, 0.48%↑) 210 1
MobileNetV2 [fp32,imagenet] (TensorFlow) little-core,full-inference with IREE-Dylib-Sync @ Pixel-4 (CPU-ARMv8.2-A) 1332 (vs. 1337, 0.37%↓) 1334 9
MobileNetV2 [fp32,imagenet] (TensorFlow) 3-thread,little-core,full-inference with IREE-VMVX @ SM-G980F (CPU-ARMv8.2-A) 62679 (vs. 62885, 0.33%↓) 62854 273
MobileNetV2 [fp32,imagenet] (TensorFlow) 1-thread,little-core,full-inference with IREE-Dylib @ Pixel-4 (CPU-ARMv8.2-A) 1330 (vs. 1334, 0.30%↓) 1332 4
MobileNetV3Small [fp32,imagenet] (TensorFlow) little-core,full-inference with IREE-Dylib-Sync @ SM-G980F (CPU-ARMv8.2-A) 374 (vs. 373, 0.27%↑) 374 1
MobileNetV2 [fp32,imagenet] (TensorFlow) little-core,full-inference with IREE-Dylib-Sync @ SM-G980F (CPU-ARMv8.2-A) 1253 (vs. 1255, 0.16%↓) 1255 5
MobileNetV2 [fp32,imagenet] (TensorFlow) 1-thread,little-core,full-inference with IREE-Dylib @ SM-G980F (CPU-ARMv8.2-A) 1255 (vs. 1256, 0.08%↓) 1256 3
MobileNetV2 [fp32,imagenet] (TensorFlow) 3-thread,little-core,full-inference with IREE-VMVX @ Pixel-4 (CPU-ARMv8.2-A) 70941 (vs. 70963, 0.03%↓) 70946 20
MobileNetV3Small [fp32,imagenet] (TensorFlow) 3-thread,little-core,full-inference with IREE-VMVX @ SM-G980F (CPU-ARMv8.2-A) 16798 (vs. 16794, 0.02%↑) 16798 14
MobileNetV3Small [fp32,imagenet] (TensorFlow) 3-thread,little-core,full-inference with IREE-VMVX @ Pixel-4 (CPU-ARMv8.2-A) 18947 (vs. 18950, 0.02%↓) 18943 20
MobileNetV3Small [fp32,imagenet] (TensorFlow) 3-thread,big-core,full-inference with IREE-Dylib @ SM-G980F (CPU-ARMv8.2-A) 79 (vs. 79, 0.00%) 79 3
MobileNetV3Small [fp32,imagenet] (TensorFlow) 1-thread,big-core,full-inference with IREE-Dylib @ SM-G980F (CPU-ARMv8.2-A) 46 (vs. 46, 0.00%) 46 0
MobileNetV3Small [fp32,imagenet] (TensorFlow) big-core,full-inference with IREE-Dylib-Sync @ SM-G980F (CPU-ARMv8.2-A) 46 (vs. 46, 0.00%) 45 1
MobileNetV2 [fp32,imagenet] (TensorFlow) 1-thread,big-core,full-inference with IREE-Dylib @ Pixel-4 (CPU-ARMv8.2-A) 170 (vs. 170, 0.00%) 170 0
MobileNetV2 [fp32,imagenet] (TensorFlow) full-inference with IREE-Vulkan @ Pixel-4 (GPU-Adreno-640) 110 (vs. 110, 0.00%) 110 0
MobileNetV3Small [fp32,imagenet] (TensorFlow) 1-thread,big-core,full-inference with IREE-Dylib @ Pixel-4 (CPU-ARMv8.2-A) 51 (vs. 51, 0.00%) 51 0
MobileNetV3Small [fp32,imagenet] (TensorFlow) big-core,full-inference with IREE-Dylib-Sync @ Pixel-4 (CPU-ARMv8.2-A) 60 (vs. 60, 0.00%) 61 0
MobileNetV3Small [fp32,imagenet] (TensorFlow) full-inference with IREE-Vulkan @ Pixel-4 (GPU-Adreno-640) 42 (vs. 42, 0.00%) 42 0
MobileNetV3Small [fp32,imagenet] (TensorFlow) kernel-execution with IREE-Vulkan @ Pixel-4 (GPU-Adreno-640) 38 (vs. 38, 0.00%) 38 1
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment