Created by: ysh329
- Enhance cpu profiler with real cpu backend kernel name；
- Enhance Conv op profiler info with
bias，act_type
arm cpu profiler for tf_mobilenetv1
===== Detailed Dispatch Profiler Summary: N/A, Exclude 1 warm-ups =====
OperatorType KerneAttr KernelName Remark InDim FilterDim OutDim Avg(ms) Min(ms) Max(ms) Last(ms) Avg(%) GOPs GOPS clAvg(ms) clMin(ms) clMax(ms) clAvg(%)
conv2d arm/float/NCHW conv_3x3s2_direct_fp32 3x3p0s2g1d1BiasRelu6 1x3x224x224 32x3x3x3 1x32x112x112 3.026 2.540 3.189 3.058 2.84% 0.02 7.16 0.000 0.000 0.000 0.00%
depthwise_conv2d arm/float/NCHW conv_depthwise_3x3_fp32 3x3p1s1g32d1BiasRelu6 1x32x112x112 32x1x3x3 1x32x112x112 1.483 1.350 1.619 1.465 1.39% 0.01 4.87 0.000 0.000 0.000 0.00%
conv2d arm/float/NCHW conv1x1s1_gemm 1x1p0s1g1d1BiasRelu6 1x32x112x112 64x32x1x1 1x64x112x112 5.623 4.934 5.921 5.687 5.28% 0.05 9.14 0.000 0.000 0.000 0.00%
depthwise_conv2d arm/float/NCHW conv_depthwise_3x3_fp32 3x3p0s2g64d1BiasRelu6 1x64x112x112 64x1x3x3 1x64x56x56 1.349 1.230 1.500 1.469 1.27% 0.00 2.68 0.000 0.000 0.000 0.00%
conv2d arm/float/NCHW conv1x1s1_gemm 1x1p0s1g1d1BiasRelu6 1x64x56x56 128x64x1x1 1x128x56x56 5.003 4.225 5.259 4.916 4.70% 0.05 10.27 0.000 0.000 0.000 0.00%
depthwise_conv2d arm/float/NCHW conv_depthwise_3x3_fp32 3x3p1s1g128d1BiasRelu6 1x128x56x56 128x1x3x3 1x128x56x56 1.263 1.137 1.395 1.304 1.19% 0.01 5.72 0.000 0.000 0.000 0.00%
conv2d arm/float/NCHW conv1x1s1_gemm 1x1p0s1g1d1BiasRelu6 1x128x56x56 128x128x1x1 1x128x56x56 9.535 8.231 10.327 9.517 8.95% 0.10 10.78 0.000 0.000 0.000 0.00%
depthwise_conv2d arm/float/NCHW conv_depthwise_3x3_fp32 3x3p0s2g128d1BiasRelu6 1x128x56x56 128x1x3x3 1x128x28x28 0.571 0.489 0.766 0.492 0.54% 0.00 3.16 0.000 0.000 0.000 0.00%
conv2d arm/float/NCHW conv1x1s1_gemm 1x1p0s1g1d1BiasRelu6 1x128x28x28 256x128x1x1 1x256x28x28 4.535 3.806 4.923 4.657 4.26% 0.05 11.33 0.000 0.000 0.000 0.00%
depthwise_conv2d arm/float/NCHW conv_depthwise_3x3_fp32 3x3p1s1g256d1BiasRelu6 1x256x28x28 256x1x3x3 1x256x28x28 0.660 0.549 0.786 0.658 0.62% 0.00 5.47 0.000 0.000 0.000 0.00%
conv2d arm/float/NCHW conv1x1s1_gemm 1x1p0s1g1d1BiasRelu6 1x256x28x28 256x256x1x1 1x256x28x28 8.671 7.166 9.332 8.595 8.14% 0.10 11.85 0.000 0.000 0.000 0.00%
depthwise_conv2d arm/float/NCHW conv_depthwise_3x3_fp32 3x3p0s2g256d1BiasRelu6 1x256x28x28 256x1x3x3 1x256x14x14 0.330 0.274 0.433 0.334 0.31% 0.00 2.73 0.000 0.000 0.000 0.00%
conv2d arm/float/NCHW conv1x1s1_gemm 1x1p0s1g1d1BiasRelu6 1x256x14x14 512x256x1x1 1x512x14x14 4.281 3.499 4.538 4.247 4.02% 0.05 12.00 0.000 0.000 0.000 0.00%
depthwise_conv2d arm/float/NCHW conv_depthwise_3x3_fp32 3x3p1s1g512d1BiasRelu6 1x512x14x14 512x1x3x3 1x512x14x14 0.479 0.394 0.624 0.479 0.45% 0.00 3.77 0.000 0.000 0.000 0.00%
conv2d arm/float/NCHW conv1x1s1_gemm 1x1p0s1g1d1BiasRelu6 1x512x14x14 512x512x1x1 1x512x14x14 8.325 7.136 9.077 8.295 7.82% 0.10 12.34 0.000 0.000 0.000 0.00%
depthwise_conv2d arm/float/NCHW conv_depthwise_3x3_fp32 3x3p1s1g512d1BiasRelu6 1x512x14x14 512x1x3x3 1x512x14x14 0.485 0.469 0.605 0.476 0.46% 0.00 3.73 0.000 0.000 0.000 0.00%
conv2d arm/float/NCHW conv1x1s1_gemm 1x1p0s1g1d1BiasRelu6 1x512x14x14 512x512x1x1 1x512x14x14 8.287 8.129 8.653 8.249 7.78% 0.10 12.40 0.000 0.000 0.000 0.00%
depthwise_conv2d arm/float/NCHW conv_depthwise_3x3_fp32 3x3p1s1g512d1BiasRelu6 1x512x14x14 512x1x3x3 1x512x14x14 0.481 0.469 0.613 0.476 0.45% 0.00 3.76 0.000 0.000 0.000 0.00%
conv2d arm/float/NCHW conv1x1s1_gemm 1x1p0s1g1d1BiasRelu6 1x512x14x14 512x512x1x1 1x512x14x14 8.639 8.422 8.929 8.814 8.11% 0.10 11.90 0.000 0.000 0.000 0.00%
depthwise_conv2d arm/float/NCHW conv_depthwise_3x3_fp32 3x3p1s1g512d1BiasRelu6 1x512x14x14 512x1x3x3 1x512x14x14 0.478 0.469 0.555 0.479 0.45% 0.00 3.78 0.000 0.000 0.000 0.00%
conv2d arm/float/NCHW conv1x1s1_gemm 1x1p0s1g1d1BiasRelu6 1x512x14x14 512x512x1x1 1x512x14x14 8.301 8.149 8.538 8.259 7.79% 0.10 12.38 0.000 0.000 0.000 0.00%
depthwise_conv2d arm/float/NCHW conv_depthwise_3x3_fp32 3x3p1s1g512d1BiasRelu6 1x512x14x14 512x1x3x3 1x512x14x14 0.482 0.470 0.605 0.477 0.45% 0.00 3.74 0.000 0.000 0.000 0.00%
conv2d arm/float/NCHW conv1x1s1_gemm 1x1p0s1g1d1BiasRelu6 1x512x14x14 512x512x1x1 1x512x14x14 8.357 8.100 11.169 8.286 7.85% 0.10 12.30 0.000 0.000 0.000 0.00%
depthwise_conv2d arm/float/NCHW conv_depthwise_3x3_fp32 3x3p0s2g512d1BiasRelu6 1x512x14x14 512x1x3x3 1x512x7x7 0.260 0.254 0.388 0.254 0.24% 0.00 1.74 0.000 0.000 0.000 0.00%
conv2d arm/float/NCHW conv1x1s1_gemm 1x1p0s1g1d1BiasRelu6 1x512x7x7 1024x512x1x1 1x1024x7x7 4.922 4.792 5.875 4.835 4.62% 0.05 10.44 0.000 0.000 0.000 0.00%
depthwise_conv2d arm/float/NCHW conv_depthwise_3x3_fp32 3x3p1s1g1024d1BiasRelu6 1x1024x7x7 1024x1x3x3 1x1024x7x7 0.335 0.326 0.448 0.331 0.31% 0.00 2.69 0.000 0.000 0.000 0.00%
conv2d arm/float/NCHW conv1x1s1_gemm 1x1p0s1g1d1BiasRelu6 1x1024x7x7 1024x1024x1x1 1x1024x7x7 9.556 9.293 11.135 9.424 8.97% 0.10 10.75 0.000 0.000 0.000 0.00%
pool2d arm/float/NCHW NotImpl avg7x7s2p0VALID 1x1024x7x7 N/A 1x1024x1x1 0.034 0.032 0.037 0.035 0.03% 0.00 1.46 0.000 0.000 0.000 0.00%
conv2d arm/float/NCHW conv1x1s1_gemm 1x1p0s1g1d1Biasunk 1x1024x1x1 1001x1024x1x1 1x1001x1x1 0.710 0.686 0.835 0.711 0.67% 0.00 2.89 0.000 0.000 0.000 0.00%
squeeze2 arm/float/NCHW NotImpl N/A 1x1001x1x1 N/A 1x1001 0.004 0.003 0.005 0.004 0.00% 0.00 0.00 0.000 0.000 0.000 0.00%
reshape2 host/any/any NotImpl N/A 1x1001 N/A 1x1001 0.004 0.002 0.008 0.004 0.00% 0.00 0.00 0.000 0.000 0.000 0.00%
softmax arm/float/NCHW NotImpl axis1 1x1001 N/A 1x1001 0.022 0.018 0.133 0.021 0.02% 0.00 0.27 0.000 0.000 0.000 0.00%
reshape2 host/any/any NotImpl N/A 1x1001 N/A 1x1001 0.002 0.001 0.003 0.002 0.00% 0.00 0.00 0.000 0.000 0.000 0.00%