Porting HPC Applications to Intels Xeon Phi Processor Architecture

Statement of the Problem or Situation that is Being Addressed: The Xeon Phi many-core processor holds great promise for enabling high performance computation on workstations and small clusters, which could accelerate the adoption of HPC modeling and simulation by small and medium enterprises (SMEs) engaged in manufacturing and engineering. An easy-to-use and robust tool chain for understanding, porting and refactoring existing HPC codes is essential to unlocking the full potential of Xeon Phi for technical computing in the manufacturing sector. Statement of How this Problem or Situation is Being Addressed: EP Analytics, Inc. is developing and commercializing a suite of software tools, PRISM (which stands for Performance Recommendations based on binary Instrumentation and Source Mining), for assisting open source developers and Independent Software Vendors (ISVs) in porting HPC modeling and simulation applications to Intels Xeon Phi processor architecture. EP Analytics will leverage over a decade of HPC tools research and development, including recent work involving the Xeon Phi architecture, to complete this work. The utility of PRISM will not be limited to standalone or small systems and will be equally useful for large-scale HPC systems or cloud-based HPC systems. What is to be done in Phase I: In Phase I of the project, EP Analytics will complete the following tasks: (1) implement a fully functional binary instrumentation (BI) toolkit for applications compiled to run in native-mode on the Xeon Phi; (2) Implement three analysis tools based on the BI toolkit (static analysis, vectorization-meter and memory-characterization); (3) harden, test, and validate a prototype source-code analysis tool; and (4) demonstrate the feasibility of our proposed methodology that provides porting and performance optimization recommendations to developers porting production applications to the Xeon Phi. Commercial Applications and Other Benefits: The development tools being created by EP Analytics for Intels Xeon Phi processor can enable high performance modeling and simulation applications across a wide range of commercial sectors including aerospace, defense, electronic design automation, life sciences, manufacturing, oil & gas, and others. Key Words: High performance computing, processors, software development, Xeon Phi, HPC, technical computing, advanced manufacturing Summary for Members of Congress: EP Analytics, Inc., of San Diego, CA (52nd Congressional District), has been awarded a grant to create software development tools for the Xeon Phi computer processor. Xeon Phi is an advanced many-core processor from Intel and can enable high performance modeling and simulation applications that can be used by manufacturers to improve product designs, time-to-market, and overall competitiveness. The software tools developed by EP Analytics will enable the implementation of advanced modeling and simulation on technical workstations and small computing clusters accessible to small- and medium-sized manufacturing and product development companies.