Early Career Lighting Talks – Running Molecular Dynamic Simulation Workloads on Oracle Cloud Infrastructure
Diversity Equity Inclusion (DEI)
Education and Training and Outreach
HPC Community Collaboration
TimeSunday, 14 November 20213:35pm - 3:40pm CST
DescriptionHPC is the last workload that will move to the cloud, and even then, some of these workloads may never will. There are several barriers to cloud adoption including cost, data movement, and security. However, there are far more benefits to running HPC workloads on the cloud. You’re able to burst, which eliminates the need to have physical servers that may limit your workloads when you need temporary capacity. The pay for only what you use model decreases costs of keeping your own data centers, eliminating the need to buy new hardware. In addition to reduced costs, you’re able to spin up clusters immediately, capacity no longer being an issue. Even with benefits like these, people need to see first-hand how their workloads will perform in the cloud. That is why I ran benchmarks for NVIDIA Clara Parabricks Pipelines, GROMACS, and NAMD to show the power of the cloud.