The Center for Hybrid Multicore Productivity Research (CHMPR)
Multicore processors — commodity chips for PCs as well as high-end clusters — have evolved from 2 to 4 cores on a chip to 64 cores on a chip. GPU accelerators with potentially thousands of cores are in PCs and clusters as hybrid configurations. Computer systems built from hundreds of thousands of multicore processors have recently cracked the Petaflop barrier. And quantum computers have become commercially available. Despite these developments in high-performance computing, significant challenges still persist before applications can effectively take advantage of these architectures.
The University of Maryland, Baltimore County (UMBC), the University of California, San Diego and the University of Utah together formed the National Science Foundation’s Center for Hybrid Multicore Productivity Research. This NSF Industry-University Cooperative Research program entered its second phase of support on Sept. 1, 2014, with an extraordinary roster of industry and government partners who are focused on addressing productivity, performance, and scalability issues in the continuous evolution of multicore architectures and open source tools applied to problems of national significance.
Multicore computational facilities, engineering talent, and academic expertise support the research problems of participating industry and government members. Collaborations in big data analytics for health IT, climate change, brain and cell imaging, genomics, interactive digital media and data intensive parallel computing paradigms have yielded new technologies, high-impact data sharing, and joint ventures.
If you or your company has an interest in becoming a member, or you have any questions or comments please contact us.