Clouds and Distributed Computing
TimeMonday, June 17th6pm - 6:20pm
DescriptionIt has become critical for many organizations to find high-performance computing (HPC) solutions that can scale with their compute-intensive workloads as a way to gain a competitive advantage. This drive for virtually unlimited HPC capacity and extreme-scale, whether it be on-premise, in the cloud or in a hybrid environment, makes sense given the right environment. Yet organizations are looking for ways to speed up innovation by having greater flexibility, predictability and scale as their HPC requirements grow and evolve.
Western Digital, a leader in data-centric solutions, alongside AWS and Univa, sought to achieve extreme-scale within its data center as a way to build a cloud-scale HPC cluster on AWS to simulate key elements of upcoming designs for their next-generation hard disk drives (HDD). The three companies worked together to evaluate the impact of running electro-magnetic engineering simulations on a massive HPC cluster built on AWS using Amazon EC2 Spot Instances, with the goal of completing the project in the smallest amount of time and at the lowest cost. As a result, Western Digital ran approximately 2.5 million simulation tasks on a Spot-based cluster of over one million vCPUs to determine optimal device characteristics that would help improve product quality, performance, reliability and durability for next-generation HDDs. The collaborative effort achieved a key milestone by completing what was initially a 480-hour (20-day) project in just 8 hours. This presentation will discuss how Western Digital was able to work with Univa and AWS to achieve extreme-scale HPC to achieve a competitive edge.