You are here

Hybrid Multicore Productivity Research (CHMPR)

University of California San Diego

University of Maryland, Baltimore County

University of Utah

Last Reviewed: (not done)

Research to optimize productivity, performance and scalability of applications of hybrid multicore processors and to investigate and prototype new applications and cloud computing services made possible by hybrid multicore computing capabilities that meet the needs of industry and government partners.

Center Mission and Rationale

The NSF Center for Hybrid Multicore Productivity Research (CHMPR) and its affiliated industry and government partners aim to be the preeminent cooperative research consortium for addressing the productivity, performance and scalability issues in meeting the insatiable computational demands of its sponsors applications through the continuous evolution of multicore architectures and open source tools. 

Multicore processors have become commodity chips for PC’s as well as high-end clusters, yet few applications effectively take advantage of these architectures. As the number of processors on a single chip is expected to increase from single digits to triple digits in the next decade, scaling industry and government applications to grow linearly in performance becomes the defining challenge for the multicore industry.  CHMPR’s mission is to advance the computational training of affiliated university students and faculty in multicore technologies; aggregate best practices for approaches to multicore systems and software development; create forum for active engagements between industry and academia; provide challenging computational research opportunities to university faculty in areas of interest to industry and government sponsors; develop and promote fundamental research in the scalability of multicore processing performance and cloud computing services; and offer courses and provide internship opportunities for students to engage with industry in projects of relevance to the environment, human health, science and socio- and econophysics, and our nation’s welfare. The Center was established as an NSF I/UCRC in 2009.

Research program

Hybrid Multicore Computing

The CHMPR research programs draws on the diverse computational science expertise available at the partnering universities to address industry’s current and future multicore application needs. The specific multicore applications and cloud computing services will be focused on the following:

  • Performance testing and scaling optimization of very large hybrid multicore systems applied to data intensive modeling and simulation;
  • S/W Tool development and algorithmic design for parallel distributed and cloud computing;
  • The application of multicore capabilities to interactive digital media and 3-D graphic environments and virtual worlds;
  • Exploration of the performance of new applications made possible by multicore capabilities such as genetic algorithms for personalized medicine, natural disaster simulation assessments and social media connectivity; and
  • Hybrid computing that hosts multicore processors as accelerators.

Facilities and Laboratory

A multidisciplinary suite of leading research laboratories include:

  • University of Maryland, Baltimore County Multicore Computational Center http://chmpr.umbc.edu/
  • University of California, San Diego A. C. Clarke Center for Human Imagination http://imagination.ucsd.edu/ ACCCHI
  • University of Utah Center for Scalable Analytics and Informatics