Image recognition training: multiple concurrent jobs

Highly efficient performance. All jobs ran at full speed even when the cluster was fully used. The NetApp storage systems delivered training performance comparable to local SSD storage while enabling easy sharing of data between servers.

Image recognition training: scale out

Highly efficient for up to four nodes. At that point, scale out was less efficient but still feasible. Using a higher-speed computational network improves scalability. The NetApp storage system delivered training performance comparable to local SSD storage while enabling easy sharing of data between servers.