Deep learning applications have been on a continuous exponential growth curve. This growth has been even more pervasive with the advent of generative AI workloads. Deep learning workloads, including generative AI workloads, require intensive compute to train. This white paper provides the performance results that customers can expect with the new generation of Dell PowerEdge servers that enable faster training of machine learning workloads. These performance numbers allow for data centers to enable better design and purchase decisions.
In this whitepaper, we outline the system performance of servers using the MLPerf Training v3.0 benchmark that was released recently. Dell servers delivered outstanding performance across different workloads such as image classification, medical image segmentation, lightweight and heavy weight object detection, speech recognition, NLP, and recommendation. Customers can refer to this data to anticipate workload acceleration to reflect their setting.