High Performance Computing Training Workshop

Category

Workshops

Published on

Abstract

HPC is playing an increasingly important role in many areas of research in Science and Engineering, and there is an urgent need to train more researchers in the use of this technology. The National Science Foundation is making large investments in HPC to enable research solutions that were not possible in the past. Interestingly, the National Science Board meeting to decide on over $300 million of spending on HPC took place at the same time as our HPC Training Workshop.

The agenda for this workshop was designed to address these needs. Topics covered included:
  • Parallel Programming Models
  • Multicore architectures
  • High-end HPC architectures
  • MPI
  • OpenMP
  • TotalView parallel debugger
  • Performance analysis
  • Condor
A new feature of the workshop was the inclusion of a presentation on multicore architectures by Prof Vijaykumar from Purdue’s ECE Dept, who is a nationally recognized expert in computer architecture. Multicore architectures present a new challenge and an opportunity in all areas of computing including HPC, and including this topic in the workshop was very timely. Other presentations by faculty members covered Parallel Programming Models (Prof. Sam Midkiff, ECE), HPC architectures (Prof Mithuna Thottethodi, ECE) and OpenMP (Prof Rudi Eigenmann, ECE). Additional topics were covered by RCAC staff members, on topics that included MPI programming, the use of the TotalView parallel debugger, the Condor system for accessing unused cycles, and the use of software tools for analyzing performance and increasing programmer productivity. image

Sponsored by

Cite this work

Researchers should cite this work as follows:

  • (2007), "High Performance Computing Training Workshop," https://nanohub.org/resources/3336.

    BibTex | EndNote

Time

Location

Envision Center, Purdue University, West Lafayette, IN

Tags

In This Workshop

  1. HPCW Condor: High Throughput Computing

    09 Oct 2007 | Online Presentations | Contributor(s): Thomas (Tom) Kesler

    Condor is a research project of the University of Wisconsin. It offers an environment called "High Throughput Computing" for compute-intensive jobs. Condor can manage a collection of compute nodes and can harness wasted CPU power from otherwise idle workstations of a cluster or desktop. Condor...

  2. HPCW Parallel Programming Models

    09 Oct 2007 | Online Presentations | Contributor(s): Sam Midkiff

  3. HPCW High-end HPC Architectures

    09 Oct 2007 | Online Presentations | Contributor(s): Mithuna Thottethodi

  4. HPCW Introduction to Parallel Programming with MPI

    05 Dec 2007 | Online Presentations | Contributor(s): David Seaman

    Single-session courseillustrating message-passing techniques. The examples include point-to-pointand collective communication using blocking and nonblocking transmission. Oneapplication illustrates the manager/worker model with buffered communications.Code examples provided in C, C++, Fortran 77,...