Loading Events
  • This event has passed.

Intel webinars on HPC available for viewing on U-M campus — Oct. 21

October 21, 2013 @ 12:00 am

A viewing room will be available on North Campus on October 21 for a series of webinars presented by Intel on how its services can be used in various aspects of high performance computing, titled “Introduction to Intel MIP and Cluster Tools.” The webinars begin at 9 a.m. and will be shown in Videoconference Room 1180 in the Duderstadt Center. Viewers can drop in for any sessions that interest them. Agenda: 9 – 10 a.m.: Section 1 — Intel Cluster Studio XE

  1. Introduction to Intel Cluter Studio XE
  2. Intel MPI Library
    • Overview
    • What’s New in the 2013 release
  3. Intel Trace Analyzer and Collector
    • Tracing your MPI application with Intel Trace Collector
    • Performance analysis with Intel Trace Analyzer
    • Ideal Interconnect Simulator & Imbalance Diagrams
    • Demo
    • MPI Correctness Checking
  4. Intel MPI Benchmarks
  5. Questions and Answers

10 a.m. – noon: Section 2 — Exploiting Parallelism for Intel Processors

  1. Parallelism Growth:
    • Rapid growth in parallelism
    • Enabling advances in parallelism
  2. Intel Compilers Key Features:
    • Vectorization, IPO, PGO, HLO, Parallel programming models (Cilk Plus)
    • Other optimization considerations

Noon – 12:30 p.m. — Lunch 12:30 – 2:30 p.m.: Section 3 — Introduction to Intel Performance and Threading Analysis tools

  1. Introduction to Intel VTune Amplifier XE profiler
    • Types of Analysis
      • Hotspot analysis
      • Concurrency Analysis
      • Locks and Waits Analysis
    • User and Synchronization API, Frame/Task Analysis
    • Command Line Interface, Installation, Remote Collection
    • Conclusion
  2. Introduction to Intel Inspector XE
    • The Inspector XE workflow and walk thru
    • Dynamic Memory and Threading Analysis
    • Static Analysis
    • Readying your sources and builds
    • Managing analysis results
    • Team collaboration
    • Advanced features

2:30 – 3:30 p.m.: Section 4 — Introduction to Intel Performance Libraries

  1. Intel Math Kernel Library (Intel MKL)
    • Latest features in forthcoming 11.0
    • Current Work
    • Intel Xeon Phi support
    • Bitwise Reproducibility
  2. Introduction to Intel Integrated Performance Primitives (Intel IPP)
    • Domains and Functions Covered
  3. Introduction to Intel Threading Building Blocks (Intel TBB)
  4. Questions and Answers

3:30 – 5 p.m.: Open Discussion / Questions and Answers

Details

Date:
October 21, 2013
Time:
12:00 am
Event Tags:
,