CHICAGO, Oct. 9, 2018 — Univa, a leading innovator in workload management and optimization solutions for high-performance computing (HPC) and enterprise customers, today announced that it is working with the Information Sciences Institute (ISI), a unit of the University of Southern California’s Viterbi School of Engineering, to help manage their growing infrastructure and accelerate the group’s machine learning research.
ISI is a world leader in research and development of advanced information processing, computer and communications technologies. ISI consists of 350 engineers, research scientists, graduate students and staff and includes machine learning/artificial intelligence, cyber security, novel electronics, HPC architectures and quantum computing. The organization’s video, image, speech and text analytics (VISTA) group spent the past three years advancing the state of research for facial recognition, which has significant implications for security and commerce. As compute workload demand increased for this group, it became evident that a more sophisticated way to distribute and manage resources among its multiple users was required. Therefore, the group’s challenge was to build an infrastructure that could scale and optimize comprehensive data simulations. They selected Univa Grid Engine to help manage their growing infrastructure, based on Univa’s enterprise-class performance and capabilities, built-in advanced GPU support, detailed documentation and ongoing product upgrades.
“The basis for artificial intelligence and machine learning research is to create neural networks to help solve problems,” said Stephen Rawls, programmer and research analyst at ISI. “But training artificial neural networks requires a lot of data and a significant amount of GPU time for tuning up parameters and running multiple sets of experiments simultaneously. We needed a reliable, powerful workload management platform that would enhance performance and could run complex, diverse workloads across multiple users. With Univa Grid Engine, we now have an infrastructure that schedules workloads to GPUs and is able to operate at 95 percent capacity with lower overall costs.”
As an example, one of the ways ISI scientists are teaching computers how to recognize faces is by extracting facial landmarks. Their system can contain 68 landmarks, such as eyebrows, nose and mouth. Using code and algorithms, the project utilizes deep learning to teach computers how to mimic the way in which neurons in the brain communicate with each other. Additionally, specific experiments often required weeks of compute time to run terabytes of data. For instance, one experiment contained the image processing of over three million images. The VISTA team used Univa Grid Engine software to set up the parallel processing and manage dependencies for the entire process without fail.
“There is no question that machine learning projects are poised for massive growth over the next two years, and ISI is at the forefront of this trajectory,” said Gary Tyreman, CEO at Univa. “Additionally, we are seeing HPC, GPUs and hybrid cloud playing more pivotal roles with machine learning projects, and we look forward to helping organizations like ISI’s VISTA group manage their growing infrastructure while taking advantage of Univa’s advanced GPU support and aggressive development roadmap.”
About Univa Corporation
Univa is the leading innovator of workload management products that optimize the performance of applications, services, and containers. Univa enables enterprises to fully utilize and scale compute resources across on-premise, cloud, and hybrid infrastructures. Advanced reporting and monitoring capabilities provide insights to make scheduling decisions and achieve even faster time-to-results. Univa’s solutions help hundreds of companies to manage thousands of applications and run billions of tasks every day. Univa is headquartered in Chicago, with offices in Canada and Germany. For more information, please visit www.univa.com.
Source: Univa Corporation