CentOS machine problems with tensorflow

Question

I have a GPU CentOS machine with OS version 7.2

I installed the tensorflow version 1.0 using pip (I did not compile)

When I run a problem using updated keras, I get

W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE3 instructions, but these are available on your machine and could speed up CPU computations.

W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.1 instructions, but these are available on your machine and could speed up CPU computations.

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.

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.

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.

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.

Any thoughts on how to fix this?

Answer 1

These warnings are harmless, but indicate that—if you wanted to—you could compile a more optimized version of TensorFlow for your local machine. The published binaries for TensorFlow are somewhat conservative in what platform-specific optimizations they apply, so that they can run on a wide range of machines. Note that these optimizations only affect CPU performance, so for GPU-accelerated models they will not have as much of an effect.

Building TensorFlow from source with the options suggested in this answer will give you a version that does not emit these warnings.