New Hires offer Personalized Service for HPC Users across Campus

Thursday, April 27, 2017
Caption: Lead IT Research Engineer Firat Coskun (far left) stands with DoIT Director of Systems Support Sanjay Kapur (4th from left) who is next to IACS Director Robert Harrison and Senior HPC Systems Engineer Eric Rosenberg, along with the new HPC student assistants. The group is standing in front of the new SeaWulf cluster. 

Research Computing and Cyberinfrastructure Staff Help PIs focus on their Science

Through a generous and transformative award of funds from Stony Brook University’s Chief Information Officer Melissa Woo, a mix of undergraduate and graduate students from multiple disciplines have been hired to leverage their skills and offer a catered experience to those on and off-campus researchers who use the University’s high-performance computer clusters LIred and SeaWulf.

Not only are individual researchers using the cluster, but so are undergraduate and graduate classes. “We’ve brought on multiple classes from Applied Mathematics, Biology, Computer Science and Tech & Society departments on to the clusters, and with the help of our students, we have been able to offer both the professors and their students much more personalized support,” said Lead IT Research Engineer Firat Coskun.

“We’ve approached all of our top users (accounting for ~70% of the computational time used across the campus clusters) and assigned one of our students to each of their groups. Already we have received positive feedback from multiple PIs about how they’re now able to focus on their science rather than spending time tracking down pesky problems with compilation or environment variables. In one case, we’ve introduced an entire department (Ecology & Evolution) to computational science via a training class one of our students conducted at the Institute for Advanced Computational Science [IACS].”

LIred is a 100-compute node system acquired as a result of a $1M grant from the New York State Department of Economic Development in 2014. Each node has two Intel Xeon E5-2690v3 CPUs, for 24 cores total, 128 Gigabytes of DDR4 Memory, and 56 Gigabit/sec InfiniBand connection, for a combined peak performance of 100 Double Precision Teraflops. The system is housed in IACS.

SeaWulf is a 164-compute node system made possible by a $1.4M National Science Foundation grant (#1531492) that was matched by $300K from internal SBU sources as well as $300K from NYSTAR. Each node has two Intel Xeon E5-2683v3 CPUs, for 28 cores total, 128 Gigabytes of DDR4 Memory, and 40 Gigabit/sec InfiniBand connection, for a combined peak performance of 240 Double Precision Teraflops. The system is housed in the Computing Center.

Both systems are available for use by the entire campus community. Requests for accounts are processed through the ticketing system that can be accessed here: https://it.stonybrook.edu/help/kb/getting-a-project-on-li-red-or-seawulf

"It's great to see that Stony Brook University researchers, as well as undergraduate and graduate courses, have benefited from the funding of student research computing and cyberinfrastructure staff. I hope that we will be able to build on these initial successes to develop a mature research computing and cyberinfrastructure program that helps the University to grow its research enterprise," said Chief Information Officer Melissa Woo.

The students who were hired (pictured above) hail from a diverse group of academic departments: Ruhul from Computer Science; Rachel from Applied Mathematics and Statistics; Janet from Computer Science; John from Materials Science and Chemical Engineering; Austin from Computer Science; and Lourdes from Mechanical Engineering. Not pictured above are Ruyi from the School of Marine and Atmospheric Sciences and Lukasz from Applied Mathematics and Statistics.