Distributed Deep Learning With Horovod And MPI

Horovod

Horovod is a distributed training framework for TensorFlow. The goal of Horovod is to make distributed Deep Learning fast and easy to use.

Why not traditional Distributed TensorFlow?

The primary motivation for this project is to make it easy to take a single-GPU TensorFlow program and successfully train it on many GPUs faster. This has two aspects:

How much modifications does one have to make to a program to make it distributed, and how easy is it to run it.
How much faster would it run in distributed mode?

Internally at Uber we found that it's much easier for people to understand an MPI model that requires minimal changes to source code than to understand how to set up regular Distributed TensorFlow.

To give some perspective on that, this commit into our fork of TF Benchmarks shows how much code can be removed if one doesn't need to worry about towers and manually averaging gradients across them, tf.Server(), tf.ClusterSpec(), tf.train.SyncReplicasOptimizer(), tf.train.replicas_device_setter() and so on. If none of these things makes sense to you - don't worry, you don't have to learn them if you use Horovod.

Install

To install Horovod:

Install Open MPI or another MPI implementation.

Steps to install Open MPI are listed here.

Install the horovod pip package.

$ pip install horovod

Concepts

Horovod core principles are based on MPI concepts such as size, rank, local rank, allreduce, allgather and broadcast.

Usage

To use Horovod, make the following additions to your program:

Run hvd.init().
Pin a server GPU to be used by this process using config.gpu_options.visible_device_list. With the typical setup of one GPU per process, this can be set to local rank. In that case, the first process on the server will be allocated the first GPU, second process will be allocated the second GPU and so forth.
Wrap optimizer in hvd.DistributedOptimizer. The distributed optimizer delegates gradient computation to the original optimizer, averages gradients using allreduce or allgather, and then applies those averaged gradients.
Add hvd.BroadcastGlobalVariablesHook(0) to broadcast initial variable states from rank 0 to all other processes. Alternatively, if you're not using MonitoredTrainingSession, you can simply execute the hvd.broadcast_global_variables op after global variables have been initialized.

Example

import tensorflow as tf
import horovod.tensorflow as hvd


# Initialize Horovod
hvd.init()

# Pin GPU to be used to process local rank (one GPU per process)
config = tf.ConfigProto()
config.gpu_options.visible_device_list = str(hvd.local_rank())

# Build model...
loss = ...
opt = tf.train.AdagradOptimizer(0.01)

# Add Horovod Distributed Optimizer
opt = hvd.DistributedOptimizer(opt)

# Add hook to broadcast variables from rank 0 to all other processes during
# initialization.
hooks = [hvd.BroadcastGlobalVariablesHook(0)]

# Make training operation
train_op = opt.minimize(loss)

# The MonitoredTrainingSession takes care of session initialization,
# restoring from a checkpoint, saving to a checkpoint, and closing when done
# or an error occurs.
with tf.train.MonitoredTrainingSession(checkpoint_dir="/tmp/train_logs",
                                       config=config,
                                       hooks=hooks) as mon_sess:
  while not mon_sess.should_stop():
    # Perform synchronous training.
    mon_sess.run(train_op)

To run on a machine with 4 GPUs:

$ mpirun -np 4 python train.py

To run on 4 machines with 4 GPUs each using Open MPI:

$ mpirun -np 16 -x LD_LIBRARY_PATH -H server1:4,server2:4,server3:4,server4:4 python train.py

Check your MPI documentation for arguments to the mpirun command on your system.

Keras

Horovod supports Keras and regular TensorFlow in similar ways. See full training simple and advanced examples

Note: You must use keras.optimizers.TFOptimizer instead of native Keras optimizers.

See a full training example here.

References

Gibiansky, A. (2017). Bringing HPC Techniques to Deep Learning. Retrieved from http://research.baidu.com/bringing-hpc-techniques-deep-learning/

manoharpalanisamy/Distributed-Deep-Learning-With-Horovod-MPI