The Cornell University Center for Advanced Computing (CAC) announced an exciting new research collaboration between ourselves, NVIDIA, Dell and MathWorks . The goal of this newly formed group is to explore the capabilities of GPU computing for data manipulation using MATLAB applications. More specifically, we want to examine how we can address the needs of researchers that have large blocks of data that they need to process in parallel.One such example is research being done at Weill Cornell Medical Center, University of Michigan Health System, and Rutgers Laboratory for Computational Imaging and Bioinformatics. They are currently using NVIDIA Tesla GPUs and MATLAB to accelerate and improve the diagnosis of cancer cells using vector quantization.*** Cancer cells under microscopeImage from Constantin Friedman, MD adn Victor Brodsky, MDIt’s a tough problem – in one typical case using high-resolution images of *** cancer cells, pathologists saw a 14X speedup using MATLAB’s built-in GPU functions. In the real world, that number means a reduction in processing time from 86.9 seconds to 5.9 seconds per image. A few seconds may not seem like much, but when you imagine that you could process thousands of images a day rather than hundreds, then it’s quite significant.The work we’ve been doing at Cornell has shown that many researchers can vastly improve their ability to achieve scientific and medical insight using the power of GPU computing, but many don’t have the expertise to learn and optimize for performance using a standard low-level programming language. Now that MATLAB provides CUDA support for GPUs, a whole new population of researchers can now benefit from the computational power of GPUs without spending a lot of time and resources doing low level optimizationsAs GPU performance testing and production runs continue at Cornell, CAC will also be developing best practices for porting MATLAB code in order to help scientific researchers get the most out of GPU Computing.This is not Cornell’s first project to give the research community more access to MATLAB enabled computational resources. We previously deployed a National Science Foundation-sponsored 512-core experimental system in partnership with Purdue University, designed to provide a bridge to high-end national resources. Over 550,000 jobs ran on the experimental resource which facilitated research, student learning, and Science Gateway applications.We’d like to thank NVIDIA, Dell and Mathworks for their support in this collaboration – Cornell is conducting this research on Dell C6100 servers with the C410x PCIe expansion chassis, which supports NVIDIA Tesla M2070 GPUs. Our focus on the use of multiple GPUs on the desktop is performed via the MathWorks Parallel Computing Toolbox, and a GPU cluster via MATLAB Distributed Computing Server.If you are a scientist interested in being considered for access to the NVDIA GPUs at CAC, please reach out to us at firstname.lastname@example.org.Original post courtesy of NVIDIADr. David Lifka Biographical Sketch David is the Director of the Cornell University Center for Advanced Computing. He is a HPC industry veteran with over twenty years experience in management and technology leadership positions at Cornell and Argonne National Laboratory. His areas of expertise include parallel job scheduling and resource management systems, data management, high throughput systems, and Web services. Lifka has a Ph.D. in Computer Science from the Illinois Institute of Technology and serves on a number of academic and corporate advisory boards. His scheduling technologies have been commercially licensed by industry and he has received a ComputerWorld/Smithsonian award for innovations in IT.