This page includes examples for Open MPI that use horovodrun
. Check your
MPI documentation for arguments to the mpirun
command on your system.
Typically one GPU will be allocated per process, so if a server has 4 GPUs,
you will run 4 processes. In horovodrun
,
the number of processes is specified with the -np
flag.
To run on a machine with 4 GPUs:
$ horovodrun -np 4 -H localhost:4 python train.py
To run on 4 machines with 4 GPUs each:
$ horovodrun -np 16 -H server1:4,server2:4,server3:4,server4:4 python train.py
You can also specify host nodes in a host file. For example:
$ cat myhostfile
aa slots=2
bb slots=2
cc slots=2
This example lists the host names (aa, bb, and cc) and how many "slots" there are for each. Slots indicate how many processes can potentially execute on a node. This format is the same as in mpirun command.
To run on hosts specified in a hostfile:
$ horovodrun -np 6 -hostfile myhostfile python train.py
Usage of horovodrun
requires one of the following:
- Open MPI >= 2.X
- Spectrum MPI
- MPICH
- OpenRTE
- Gloo
- Intel(R) MPI
If you do not have MPI installed, you can run horovodrun
using Gloo. Gloo dependencies come with Horovod
automatically, and only require CMake to be available on your system at the time you install Horovod.
If you wish to use a different version of MPI, you may still be able to run Horovod using mpirun <mpi.rst> directly.
The host where horovodrun
is executed must be able to SSH to all other
hosts without any prompts.
If horovodrun
fails with a permission error, verify that you can ssh to
every other server without entering a password or
answering questions like this:
The authenticity of host '<hostname> (<ip address>)' can't be established.
RSA key fingerprint is xx:xx:xx:xx:xx:xx:xx:xx:xx:xx:xx:xx:xx:xx:xx:xx.
Are you sure you want to continue connecting (yes/no)?
To learn more about setting up passwordless authentication, see this page.
To avoid The authenticity of host '<hostname> (<ip address>)' can't be
established
prompts, add all the hosts to
the ~/.ssh/known_hosts
file using ssh-keyscan
:
$ ssh-keyscan -t rsa,dsa server1 server2 > ~/.ssh/known_hosts
In some advanced cases you might want fine-grained control over options passed to Open MPI. To learn how to run Horovod training directly using Open MPI, read Run Horovod with Open MPI.
horovodrun
automatically converts some parameters to the format supported by Intel(R) MPI mpirun
. The set of allowed options includes -np
, -H
and
ssh arguments (-p, -i). Intel(R) MPI mpirun
does not support MCA parameters, but you can set some of the options via environment variables.
For additional information refer to Intel(R) MPI official documentation.