Intel Xeon Phi coprocessor support by software tools

A number of tool vendors have announced they will be providing versions of their software tailored to supporting MIC architecture, starting with the Intel Xeon Phi co-processor.

Here are the ones I know about and can share (there are more who will make their own announcement in the future):

Compilers

  • CAPS HMPP -
    “Intel® Xeon Phi™ coprocessor provide the next step beyond acceleration in providing the HPC users ultimate flexibility in their computing need.  Our HMPP Workbench supports  Xeon Phi through our OpenHMPP directives and is soon available in beta” (September 2012)
    Francois Bodin, CTO, CAPS Entreprise
  • Intel Parallel Studio XE (available worldwide through resellers as of September 5)
  • Intel Cluster Studio XE
  • ISPC - available now

Libraries

  • IMSL Library, Rogue Wave
  • NAG Libraries, NAG Ltd.
  • Totalview debugger, Rogue Wave
  • Magma, U of Tenn.
    “Dense linear algebra is a cornerstone for many HPC applications and MAGMA. Optimized algorithms and frameworks are what we focus on and we are excited about the ease of programming and performance of the Intel® Xeon Phi™ family as we move forward in our development efforts to support it within MAGMA.”
    Jack Dongarra, University Distinguished Professor, Director – Innovative Computing Laboratory, Director – Center for Information Technology Research, University of Tennessee
  • SWARM
    "ETI's SWARM technology is a performance accelerator for multi-core/multi-node systems. SWARM is a dynamic adaptive runtime system that minimizes user exposure to physical parallelism and system complexity. It is designed to enable programming on many-core architectures by utilizing a dynamic, message-driven model of execution instead of the static scheduling and sequential computing method of conventional programming model.  We are pleased to announce that Intel's Xeon Xeon Phi™ coprocessors will be supported utilizing the SWARM technology adding to the breath of our solution for multi-core/multi-node systems.“
    Rishi Khan, Vice President of Research and Development, ET International
  • FEA structural analysis solver RADIOSS
    “Our close collaboration with Intel in support of Intel® Xeon Phi™ coprocessors has enabled us to better prepare our customers, who are excited to start working with the new technology within their clusters. Early adopters can more efficiently manage their cluster’s workload with Altair’s PBS Professional, ensuring they utilize Intel® Xeon Phi™ to its maximum benefit. We are also porting our leading FEA structural analysis solver RADIOSS to Intel® Xeon Phi™ so our CAE customers will be able to take advantage of the performance gains from this new Intel technology.”
    Bill Nitzberg, CTO, PBS Works

Debuggers

  • Rogue Wave TotalView
  • Allinea DDT
  • gdb

Tuning

  • ParaTools, TAU Performance System® - coming soon
    “With the advancement of manycore processor architectures and their realization in next-generation products comes the key need for tools that can identify areas for tuning for maximum performance.  ParaTools, Inc. (in association with the TAU Project at the University of Oregon) is please to announce support for the Intel® Xeon Xeon Phi™ coprocessors in the TAU Performance System®.”
    ParaTools, Inc., Prof. Allen D. Malony, Professor, Dept. of CIS and Director, NIC, U. Oregon, CEO and Director, ParaTools, Inc.

Virtualization

  • ScaleMP vSMP Foundation - find here information about how to obtain and use vSMP Foundation virtualization framework for MIC.
    "vSMP Foundation for Intel Xeon Phi coprocessor will virtualize both the Intel Xeon processor cores and the Intel Xeon Phi coprocessor cores to act like a single Symmetric Multiprocessing (SMP) system – sharing execution context and memory address space. Users will have easy access to all of the computing and memory resources, greatly simplifying their development and production environments.  Our beta program will be open for customers in the next 90 days.”
    Michael Schulman, Director of Marketing, ScaleMP
  • Xen is available -- more details TBD

Cluster and Workload Management

  • Altair PBS Pro - coming soon
    “Our close collaboration with Intel in support of Intel® Xeon Phi™ coprocessors has enabled us to better prepare our customers, who are excited to start working with the new technology within their clusters. Early adopters can more efficiently manage their cluster’s workload with Altair’s PBS Professional, ensuring they utilize Intel® Xeon Phi™ to its maximum benefit. We are also porting our leading FEA structural analysis solver RADIOSS to Intel® Xeon Phi™ so our CAE customers will be able to take advantage of the performance gains from this new Intel technology.”
    Bill Nitzberg, CTO, PBS Works
  • IBM Platform LSF (formerly Platform Computing LSF) and IBM Platform HPC
    “IBM is continuing our relationship with Intel through the integration of IBM Platform Computing management software and the Intel Xeon Phi architecture. IBM Platform LSF and IBM Platform HPC users will be able to intelligently allocate application workload to Intel Xeon Phi, leveraging its many core capabilities to accelerate performance.”
    Dr. William Lu, Global Product Manager, IBM Platform Computing.
  • SLURM



Of course... Intel Parallel Studio XE 2013 and Intel Cluster Studio XE 2013 as well.

There were also countless applications, many open source, that have been recompiled for MIC architecture and were being discussed around the show. Some I remember are NWChem, ENZO, ELK, MADNESS, MPI, GA, and Python... and I know I forget quite a few. Of course, Linux has been ported (and is running on both Knights Ferry and Knights Corner).

For additional tools announcements, please let me know, or post a comment.

For more complete information about compiler optimizations, see our Optimization Notice.

Comments