Computationally intensive research is supported through the operation of a scalable high performance computing system based on Dell PowerEdge servers with Intel processors and an open-sourced, Linux-based operating system. The HPC system has grown significantly since its inception in 2005 and now provides more than 600 faculty, staff and students with the ability to run parallel computational tasks that require hundreds of simultaneous processors. Internal network connectivity has been updated with a 10 gigabyte ethernet connection to ensure that parallel jobs can run without bandwidth issues. The system has the online, high-speed storage capacity of about 100 terabytes with an additional 720 terabytes of near-line storage. Currently, the overall computational performance is about 28.4 trillion floating-point operations per second (teraflops) coming from 2,792 CPU cores and another 66.8 teraflops from 25,280 GPU cores.
These are some projects that were supported by the HPC system:
Research Technologies worked with Doctoral Student Mary Ann O'Riordan who received her Ph.D. in 2013 in epidemiology and biostatistics. For her dissertation, she worked on an assessment of the safety of the anesthetic Propofol for procedural sedation from observational data using propensity score matching, a statistical matching technique that attempts to estimate the effect of a treatment, policy or other intervention by accounting for the variables that predict receiving the treatment.
A data set for a group of hypothetical patients that received no drugs during a non-operating-room procedure was created through thousands of simulations using an open-source statistical software package installed in ITS’ high performance computing system. All the simulations were run simultaneously in seconds by utilizing thousands of HPC processors, and the outputs were used for statistical analysis.
Dr. O’Riordan says, "I was fortunate to be able to benefit from the expertise of ITS’ Research Technologies staff in the recent completion of my Ph.D. dissertation. They helped me select appropriate software for a simulation portion of my work and checked my code for errors. In addition, they ran thousands of simulations simultaneously utilizing thousands of processors on the Case Western Reserve high performance computing system. This assistance saved me precious time and gave me confidence in the work that I would not have had on my own. I am sincerely grateful."
In fiscal year 2013, $252 million in research funds were awarded to Case Western Reserve faculty who utilized ITS’ Research Technology services. These faculty elected to demonstrate their appreciation at having an on-campus HPC system by investing a portion of their research funds into the purchase of equipment or memberships to the HPC system.
In addition to direct investment, many faculty members have engaged ITS to help with pre-award activities, including solution design, budget assistance, project planning and grant writing. Research Technologies staff actively participated in several large fiscal year 2013 National Institutes of Health and National Science Foundation proposals in the areas of big data research, mass storage for biomedical next generation sequencing and ultra-speed networks for science.
In fiscal year 2013, Research Technologies participated in three faculty recruitment efforts -- two in Engineering and one in the School of Medicine -- by providing consulting, expertise in engineering solutions, and service delivery to enable the new faculty to quickly implement research programs upon joining the university. ITS provided storage and compute capacity, data migration from their prior facilities and expertise in system administration that alleviated the need to create a computing infrastructure within each individual lab.
Dr. Phoebe Stewart, pharmacology professor for the School of Medicine and director of the Cleveland Center for Membrane and Structural Biology, says, “My move to Case Western Reserve was facilitated by the high performance computing and large-scale research storage resources that were already in place. The capable assistance provided by ITS’ Research Technologies enabled us to quickly and painlessly move 30 terabytes of archived data from Vanderbilt University to Case Western Reserve. Their skillful services in installing multiple image processing software packages on the HPC system also greatly eased the transition of moving my lab to Case Western Reserve with minimal loss of productivity.”
Research ShowCASE, a venue where hundreds of researchers, scientists, scholars, students, faculty, staff, alumni, business and industry leaders, and the community came together to share a day of collaboration and exhibits of real-world research applications that highlighted the full range of faculty, postdoctoral and graduate research at Case Western Reserve.
The Research Technologies team staffed a booth to share information on its research support services and to network with researchers. These conversations led to a collaboration with Dr. Satya Sahoo and his research into the development and application of computer science methods and technologies for clinical big data.
With regard to his experience with ITS, Dr. Sahoo says, “The Cloudwave platform uses the High Performance Computing system at Case Western Reserve University and relies on close technical collaborations with the Research Technologies team.”