An Introduction into Performance Analysis for HPC Systems with Open|Speedshop

Date and Time: 
2013 Thursday, April 4, and Friday, April 5
CG1-1210 South Auditorium
Jim Galarowicz and Martin Schulz

Open|SpeedShop is an open source multi platform Linux performance tool which is targeted to support performance analysis of applications running on both single node and large scale IA64, IA32, EM64T, AMD64, IBM Power PC, Cray, and Blue Gene platforms.  It supports comprehensive performance analysis for sequential, multi-threaded, and MPI applications with no need to recompile the user’s application.  There are four user interface options: batch, command line interface (CLI), graphical user interface (GUI), and a Python scripting API.  Open|SpeedShop is explicitly designed with usability in mind and is for application developers and computer scientists. The base functionality include:

  • Sampling Experiments
  • Support for Callstack Analysis
  • Hardware Performance Counters
  • MPI Profiling and Tracing
  • I/O Profiling and Tracing
  • Floating Point Exception Analysis

This tutorial will be a hands on, bring your own code sessions.  There will be short lectures and demonstrations of Open|SpeedShop along with time allocated for trying Open|SpeedShop on the tutorial participants own code.


Speaker Description: 

Jim Galarowicz has been involved with high performance computer (HPC) systems software development at Sperry Univac (compiler development), Cray Research, Inc. (compiler development and performance tool development), Silicon Graphics (debugger and performance tool development), and more recently at The Krell Institute. Currently, he is leading the Open|SpeedShop performance tools project and is primarily responsible for managing the project, design and development of portions of the Open|SpeedShop performance tool, writing project documentation and interfacing with project stakeholders. Prior to the current assignment, he managed the Silicon Graphics software development tools group between 1999 and 2006.

Dr. Martin Schulz is a Computer Scientist at the Center for Applied Scientific Computing (CASC) at Lawrence Livermore National Laboratory (LLNL). He earned his Doctorate in Computer Science in 2001 from the Technische Universität München (Munich, Germany) and also holds a Master of Science in Computer Science from the University of Illinois at Urbana Champaign. He has published over 140 peer-reviewed papers. He is the PI for the ASC/CCE project on Open|SpeedShop, and the LLNL PI for the OASCR PetaTools project on "Building a Community Tool Infrastructure around Open|SpeedShop". He is further involved in the ASCR Co-Design Centers CESAR and ExMatEx.

Event Category: