Sai Devulapalli, Dell Technologies; Tom Henson, Dell Technologies
gtc-dc 2019
We’ll discuss the challenges resulting from the mainstream adoption of machine learning and deep learning projects. As projects leave small-scale sandbox and proof of concept environments for platforms supporting large-scale production deployments of AI applications, there are several architectural considerations. We’ll demonstrate how to eliminate I/O bottlenecks so that GPUs are saturated with data; adjust data gravity, data scaling, and data economics to support petabyte-sized datasets; simplify data management; and minimize the business risks and life cycle costs of large-scale AI platforms. Platforms meeting these requirements will be most beneficial for companies. We’ll conclude by introducing the new Dell EMC and NVIDIA joint solution portfolio for large-scale machine learning and deep learning.