Execution speed is much slower than CPU in MacBook Pro?

[elife33]

In my MacBook Pro(Mid 2015), Execution speed in AMD Radeon R9 M370X Compute Engine is much slower than CPU(2.5 GHz Intel Core i7). Is this normal?

[elife33]

GPU:

• python multilayer_perceptron.py Extracting /tmp/data/train-images-idx3-ubyte.gz Extracting /tmp/data/train-labels-idx1-ubyte.gz Extracting /tmp/data/t10k-images-idx3-ubyte.gz Extracting /tmp/data/t10k-labels-idx1-ubyte.gz OpenCL platform: Apple OpenCL device: AMD Radeon R9 M370X Compute Engine I tensorflow/core/common_runtime/gpu/gpu_device.cc:989] Found device 0 with properties: name: AMD Radeon R9 M370X Compute Engine major: -1 minor: -1 memoryClockRate (GHz) 800 pciBusID 0000.0000 Total memory: 2.00GiB Free memory: 512.00MiB I tensorflow/core/common_runtime/gpu/gpu_device.cc:877] cannot enable peer access from device ordinal 0 to device ordinal 0 I tensorflow/core/common_runtime/gpu/gpu_device.cc:1011] DMA: 0 I tensorflow/core/common_runtime/gpu/gpu_device.cc:1021] 0: N I tensorflow/core/common_runtime/gpu/gpu_device.cc:1083] Creating TensorFlow device (/gpu:0) -> (device: 0, name: AMD Radeon R9 M370X Compute Engine, pci bus id: 0000.0000) cl_driver DeviceAllocate 327155712 Epoch: 0001 cost= 168.817177381 Epoch: 0002 cost= 41.912425166 Epoch: 0003 cost= 26.223643868 Optimization Finished! Accuracy: 0.913713 epoch_times [14.012919187545776, 12.755515336990356, 12.597593069076538] average_epoch_times= 12.676554203 kernel_compile_time 1.33636498451

real 0m42.142s user 0m31.954s sys 0m18.096s

CPU: QiangdeMacBook-Pro:3_NeuralNetworks elife$ time python multilayer_perceptron.py Extracting /tmp/data/train-images-idx3-ubyte.gz Extracting /tmp/data/train-labels-idx1-ubyte.gz Extracting /tmp/data/t10k-images-idx3-ubyte.gz Extracting /tmp/data/t10k-labels-idx1-ubyte.gz 2017-10-16 15:53:15.945534: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.2 instructions, but these are available on your machine and could speed up CPU computations. 2017-10-16 15:53:15.945553: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX instructions, but these are available on your machine and could speed up CPU computations. 2017-10-16 15:53:15.945568: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX2 instructions, but these are available on your machine and could speed up CPU computations. 2017-10-16 15:53:15.945572: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use FMA instructions, but these are available on your machine and could speed up CPU computations. Epoch: 0001 cost=333.116085219 Epoch: 0002 cost=97.091073028 Epoch: 0003 cost=71.633393175 Epoch: 0004 cost=58.102855900 Epoch: 0005 cost=48.550296390 Epoch: 0006 cost=43.879424709 Epoch: 0007 cost=38.416489090 Epoch: 0008 cost=35.230118899 Epoch: 0009 cost=32.181227765 Epoch: 0010 cost=30.920996809 Epoch: 0011 cost=28.272015703 Epoch: 0012 cost=26.596986358 Epoch: 0013 cost=25.029801500 Epoch: 0014 cost=23.558648788 Epoch: 0015 cost=22.414703977 Optimization Finished! Accuracy: 0.8971

real 0m32.758s user 2m15.099s sys 0m10.327s

[elife33]

Bitcoin mining Data courtesy CompuBench Radeon R9 M370X Mac 111.14 mHash/s

CompuBench 1.5 (Bitcoin mining) Data courtesy CompuBench Core i7 4870HQ 30.62 mHash/s

[elife33]

Face detection Data courtesy CompuBench Radeon R9 M370X Mac 25.65 mPixels/s

CompuBench 1.5 (Face detection) Core i7 4870HQ 18.08 mPixels/s

[hughperkins]

It is. Since all convolutions are running on the CPU currently. I never quite got around to attaching https://github.com/hughperkins/coriander-dnn to tf-coriander, so convolutions are on the cpu currently. If someone has a moment, would be good to get that working. Happy to provide guidance, meet in Google Hangouts etc, if anyone wants to take a look at that.

[datatalking]

@hughperkins, I'm open to try as i'm working to get a 560 running on high sierra via gpu