White Papers

Deep Learning Performance: Scale-up vs Scale-out
Architectures & Technologies Dell EMC | Infrastructure Solutions Group
20
The tests were run in docker environment, Figure 12 shows the different logical layers involved
in the software stack configuration. Each server is connected to the InfiniBand switch; has
installed on the Host the Mellanox OFED for Ubuntu, the Docker CE, and the GPUDirect RDMA
API; and the container image that was built with Horovod and Mellanox OFED among other
supporting libraries. To build the extended container image, we used the Horovod docker file and
modified it by adding the installation for Mellanox OFED drivers.
Figure 12: Servers Logical Design. Source: Image adapted from
https://community.mellanox.com/docs/DOC-2971
In Figure 13 below shows how PowerEdge C4130/C4140 is conncted via InifniBand fabric for
multi-node testing.
Figure 13: Using Mellanox CX5 InfiniBand adapter to connect C4130/PowerEdge C4140 in multi-node
configuration