How to compile Tensorflow with SSE4.2 and AVX instructions?

This is the message received from running a script to check if Tensorflow is working:

I tensorflow/stream_executor/dso_loader.cc:125] successfully opened CUDA library libcublas.so.8.0 locally
I tensorflow/stream_executor/dso_loader.cc:125] successfully opened CUDA library libcudnn.so.5 locally
I tensorflow/stream_executor/dso_loader.cc:125] successfully opened CUDA library libcufft.so.8.0 locally
I tensorflow/stream_executor/dso_loader.cc:125] successfully opened CUDA library libcuda.so.1 locally
I tensorflow/stream_executor/dso_loader.cc:125] successfully opened CUDA library libcurand.so.8.0 locally
W tensorflow/core/platform/cpu_feature_guard.cc:95] 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:95] The TensorFlow library wasn't compiled to use AVX instructions, but these are available on your machine and could speed up CPU computations.
I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:910] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero

I noticed that it has mentioned SSE4.2 and AVX,

1. What are SSE4.2 and AVX?
2. How do these SSE4.2 and AVX improve CPU computations for Tensorflow tasks.
3. How to make Tensorflow compile using the two libraries?

I just ran into this same problem, it seems like Yaroslav Bulatov's suggestion doesn't cover SSE4.2 support, adding --copt=-msse4.2 would suffice. In the end, I successfully built with

bazel build -c opt --copt=-mavx --copt=-mavx2 --copt=-mfma --copt=-mfpmath=both --copt=-msse4.2 --config=cuda -k //tensorflow/tools/pip_package:build_pip_package

without getting any warning or errors.

Probably the best choice for any system is:

bazel build -c opt --copt=-march=native --copt=-mfpmath=both --config=cuda -k //tensorflow/tools/pip_package:build_pip_package

(Update: the build scripts may be eating -march=native, possibly because it contains an =.)

-mfpmath=both only works with gcc, not clang. -mfpmath=sse is probably just as good, if not better, and is the default for x86-64. 32-bit builds default to -mfpmath=387, so changing that will help for 32-bit. (But if you want high-performance for number crunching, you should build 64-bit binaries.)

I'm not sure what TensorFlow's default for -O2 or -O3 is. gcc -O3 enables full optimization including auto-vectorization, but that sometimes can make code slower.

---

What this does: --copt for bazel build passes an option directly to gcc for compiling C and C++ files (but not linking, so you need a different option for cross-file link-time-optimization)

x86-64 gcc defaults to using only SSE2 or older SIMD instructions, so you can run the binaries on any x86-64 system. (See https://gcc.gnu.org/onlinedocs/gcc/x86-Options.html). That's not what you want. You want to make a binary that takes advantage of all the instructions your CPU can run, because you're only running this binary on the system where you built it.

-march=native enables all the options your CPU supports, so it makes -mavx512f -mavx2 -mavx -mfma -msse4.2 redundant. (Also, -mavx2 already enables -mavx and -msse4.2, so Yaroslav's command should have been fine). Also if you're using a CPU that doesn't support one of these options (like FMA), using -mfma would make a binary that faults with illegal instructions.

TensorFlow's ./configure defaults to enabling -march=native, so using that should avoid needing to specify compiler options manually.

-march=native enables -mtune=native, so it optimizes for your CPU for things like which sequence of AVX instructions is best for unaligned loads.

This all applies to gcc, clang, or ICC. (For ICC, you can use -xHOST instead of -march=native.)

Let's start with the explanation of why do you see these warnings in the first place.

---

Most probably you have not installed TF from source and instead of it used something like pip install tensorflow. That means that you installed pre-built (by someone else) binaries which were not optimized for your architecture. And these warnings tell you exactly this: something is available on your architecture, but it will not be used because the binary was not compiled with it. Here is the part from documentation.

> TensorFlow checks on startup whether it has been compiled with the > optimizations available on the CPU. If the optimizations are not > included, TensorFlow will emit warnings, e.g. AVX, AVX2, and FMA > instructions not included.

Good thing is that most probably you just want to learn/experiment with TF so everything will work properly and you should not worry about it

---

What are SSE4.2 and AVX?

Wikipedia has a good explanation about SSE4.2 and AVX. This knowledge is not required to be good at machine-learning. You may think about them as a set of some additional instructions for a computer to use multiple data points against a single instruction to perform operations which may be naturally parallelized (for example adding two arrays).

Both SSE and AVX are implementation of an abstract idea of SIMD (Single instruction, multiple data), which is

> a class of parallel computers in Flynn's taxonomy. It describes > computers with multiple processing elements that perform the same > operation on multiple data points simultaneously. Thus, such machines > exploit data level parallelism, but not concurrency: there are > simultaneous (parallel) computations, but only a single process > (instruction) at a given moment

This is enough to answer your next question.

---

How do these SSE4.2 and AVX improve CPU computations for TF tasks

They allow a more efficient computation of various vector (matrix/tensor) operations. You can read more in these slides

---

How to make Tensorflow compile using the two libraries?

You need to have a binary which was compiled to take advantage of these instructions. The easiest way is to compile it yourself. As Mike and Yaroslav suggested, you can use the following bazel command

bazel build -c opt --copt=-mavx --copt=-mavx2 --copt=-mfma --copt=-mfpmath=both --copt=-msse4.2 --config=cuda -k //tensorflow/tools/pip_package:build_pip_package

Let me answer your 3rd question first:

If you want to run a self-compiled version within a conda-env, you can. These are the general instructions I run to get tensorflow to install on my system with additional instructions. Note: This build was for an AMD A10-7850 build (check your CPU for what instructions are supported...it may differ) running Ubuntu 16.04 LTS. I use Python 3.5 within my conda-env. Credit goes to the tensorflow source install page and the answers provided above.

git clone https://github.com/tensorflow/tensorflow 
# Install Bazel
# https://bazel.build/versions/master/docs/install.html
sudo apt-get install python3-numpy python3-dev python3-pip python3-wheel
# Create your virtual env with conda.
source activate YOUR_ENV
pip install six numpy wheel, packaging, appdir
# Follow the configure instructions at:
# https://www.tensorflow.org/install/install_sources
# Build your build like below. Note: Check what instructions your CPU 
# support. Also. If resources are limited consider adding the following 
# tag --local_resources 2048,.5,1.0 . This will limit how much ram many
# local resources are used but will increase time to compile.
bazel build -c opt --copt=-mavx --copt=-msse4.1 --copt=-msse4.2  -k //tensorflow/tools/pip_package:build_pip_package
# Create the wheel like so:
bazel-bin/tensorflow/tools/pip_package/build_pip_package /tmp/tensorflow_pkg
# Inside your conda env:
pip install /tmp/tensorflow_pkg/NAME_OF_WHEEL.whl
# Then install the rest of your stack
pip install keras jupyter etc. etc.

As to your 2nd question:

A self-compiled version with optimizations are well worth the effort in my opinion. On my particular setup, calculations that used to take 560-600 seconds now only take about 300 seconds! Although the exact numbers will vary, I think you can expect about a 35-50% speed increase in general on your particular setup.

Lastly your 1st question:

A lot of the answers have been provided above already. To summarize: AVX, SSE4.1, SSE4.2, MFA are different kinds of extended instruction sets on X86 CPUs. Many contain optimized instructions for processing matrix or vector operations.

I will highlight my own misconception to hopefully save you some time: It's not that SSE4.2 is a newer version of instructions superseding SSE4.1. SSE4 = SSE4.1 (a set of 47 instructions) + SSE4.2 (a set of 7 instructions).

In the context of tensorflow compilation, if you computer supports AVX2 and AVX, and SSE4.1 and SSE4.2, you should put those optimizing flags in for all. Don't do like I did and just go with SSE4.2 thinking that it's newer and should superseed SSE4.1. That's clearly WRONG! I had to recompile because of that which cost me a good 40 minutes.

These are SIMD vector processing instruction sets.

Using vector instructions is faster for many tasks; machine learning is such a task.

Quoting the tensorflow installation docs:

> To be compatible with as wide a range of machines as possible, TensorFlow defaults to only using SSE4.1 SIMD instructions on x86 machines. Most modern PCs and Macs support more advanced instructions, so if you're building a binary that you'll only be running on your own machine, you can enable these by using --copt=-march=native in your bazel build command.

Thanks to all this replies + some trial and errors, I managed to install it on a Mac with clang. So just sharing my solution in case it is useful to someone.

1. Follow the instructions on Documentation - Installing TensorFlow from Sources

2. When prompted for

   Please specify optimization flags to use during compilation when bazel option "--config=opt" is specified [Default is -march=native]

then copy-paste this string:

-mavx -mavx2 -mfma -msse4.2

(The default option caused errors, so did some of the other flags. I got no errors with the above flags. BTW I replied n to all the other questions)

After installing, I verify a ~2x to 2.5x speedup when training deep models with respect to another installation based on the default wheels - Installing TensorFlow on macOS

Hope it helps

I have recently installed it from source and bellow are all the steps needed to install it from source with the mentioned instructions available.

Other answers already describe why those messages are shown. My answer gives a step-by-step on how to isnstall, which may help people struglling on the actual installation as I did.

1. Install Bazel

Download it from one of their available releases, for example 0.5.2. Extract it, go into the directory and configure it: bash ./compile.sh. Copy the executable to /usr/local/bin: sudo cp ./output/bazel /usr/local/bin

2. Install Tensorflow

Clone tensorflow: git clone https://github.com/tensorflow/tensorflow.git Go to the cloned directory to configure it: ./configure

It will prompt you with several questions, bellow I have suggested the response to each of the questions, you can, of course, choose your own responses upon as you prefer:

Using python library path: /usr/local/lib/python2.7/dist-packages
Do you wish to build TensorFlow with MKL support? [y/N] y
MKL support will be enabled for TensorFlow
Do you wish to download MKL LIB from the web? [Y/n] Y
Please specify optimization flags to use during compilation when bazel option "--config=opt" is specified [Default is -march=native]: 
Do you wish to use jemalloc as the malloc implementation? [Y/n] n
jemalloc disabled
Do you wish to build TensorFlow with Google Cloud Platform support? [y/N] N
No Google Cloud Platform support will be enabled for TensorFlow
Do you wish to build TensorFlow with Hadoop File System support? [y/N] N
No Hadoop File System support will be enabled for TensorFlow
Do you wish to build TensorFlow with the XLA just-in-time compiler (experimental)? [y/N] N
No XLA JIT support will be enabled for TensorFlow
Do you wish to build TensorFlow with VERBS support? [y/N] N
No VERBS support will be enabled for TensorFlow
Do you wish to build TensorFlow with OpenCL support? [y/N] N
No OpenCL support will be enabled for TensorFlow
Do you wish to build TensorFlow with CUDA support? [y/N] N
No CUDA support will be enabled for TensorFlow

3. The pip package. To build it you have to describe which instructions you want (you know, those Tensorflow informed you are missing).

Build pip script: bazel build -c opt --copt=-mavx --copt=-mavx2 --copt=-mfma --copt=-mfpmath=both --copt=-msse4.1 --copt=-msse4.2 -k //tensorflow/tools/pip_package:build_pip_package

Build pip package: bazel-bin/tensorflow/tools/pip_package/build_pip_package /tmp/tensorflow_pkg

Install Tensorflow pip package you just built: sudo pip install /tmp/tensorflow_pkg/tensorflow-1.2.1-cp27-cp27mu-linux_x86_64.whl

Now next time you start up Tensorflow it will not complain anymore about missing instructions.

This is the simplest method. Only one step.

It has significant impact on speed. In my case, time taken for a training step almost halved.

Refer custom builds of tensorflow

I compiled a small Bash script for Mac (easily can be ported to Linux) to retrieve all CPU features and apply some of them to build TF. Im on TF master and use kinda often (couple times in a month).

https://gist.github.com/venik/9ba962c8b301b0e21f99884cbd35082f

To compile TensorFlow with SSE4.2 and AVX, you can use directly

> bazel build --config=mkl > --config="opt" > --copt="-march=broadwell" > --copt="-O3" > //tensorflow/tools/pip_package:build_pip_package

Source: https://github.com/tensorflow/tensorflow/blob/master/tensorflow/tools/docker/Dockerfile.devel-cpu-mkl

2.0 COMPATIBLE SOLUTION:

Execute the below commands in Terminal (Linux/MacOS) or in Command Prompt (Windows) to install Tensorflow 2.0 using Bazel:

git clone https://github.com/tensorflow/tensorflow.git
cd tensorflow

#The repo defaults to the master development branch. You can also checkout a release branch to build:
git checkout r2.0

#Configure the Build => Use the Below line for Windows Machine
python ./configure.py 

#Configure the Build => Use the Below line for Linux/MacOS Machine
./configure
#This script prompts you for the location of TensorFlow dependencies and asks for additional build configuration options. 

#Build Tensorflow package

#CPU support
bazel build --config=opt //tensorflow/tools/pip_package:build_pip_package 

#GPU support
bazel build --config=opt --config=cuda --define=no_tensorflow_py_deps=true //tensorflow/tools/pip_package:build_pip_package

When building TensorFlow from source, you'll run the configure script. One of the questions that the configure script asks is as follows:

Please specify optimization flags to use during compilation when bazel option "--config=opt" is specified [Default is -march=native]

The configure script will attach the flag(s) you specify to the bazel command that builds the TensorFlow pip package. Broadly speaking, you can respond to this prompt in one of two ways:

• If you are building TensorFlow on the same type of CPU type as the one on which you'll run TensorFlow, then you should accept the default (-march=native). This option will optimize the generated code for your machine's CPU type.
• If you are building TensorFlow on one CPU type but will run TensorFlow on a different CPU type, then consider supplying a more specific optimization flag as described in the gcc documentation.

After configuring TensorFlow as described in the preceding bulleted list, you should be able to build TensorFlow fully optimized for the target CPU just by adding the --config=opt flag to any bazel command you are running.

To hide those warnings, you could do this before your actual code.

import os
os.environ['TF_CPP_MIN_LOG_LEVEL']='2'
import tensorflow as tf