Hands-on Practical Hybrid Parallel Application Performance Engineering
September 23, 2012
This tutorial presents state-of-the-art performance tools for leading-edge HPC systems founded on the Score-P community instrumentation and measurement infrastructure, demonstrating how they can be used for performance engineering of effective scientific applications based on standard MPI or OpenMP and now common mixed-mode hybrid parallelizations. Parallel performance evaluation tools from the Virtual Institute -- High Productivity Supercomputing (VI-HPS) are introduced and featured in hands-on exercises with Periscope, Scalasca, Vampir and TAU. We cover all aspects of performance engineering practice, including instrumentation, measurement (profiling and tracing, timing and hardware counters), data storage, analysis and visualization. Emphasis is placed on how tools are used in combination for identifying performance problems and investigating optimization alternatives, illustrated with a case study using a major application code.
Attendees are encouraged to bring their notebook computers and be prepared for hands-on exercises with a provided Linux Live-ISO image booted natively from DVD/USB or running within a virtual machine (e.g., VirtualBox). Due to limited time and network bandwidth available during the tutorial, those who intend to install a virtual machine on their notebook computers should do so in advance.
The "Sep12.6.64bit" 4GB ISO disk image and 5GB OVA archive that have been prepared for the tutorial are now available for advance download via the webpage http://www.vi-hps.org/training/livedvd/. They are only available as 64-bit versions, requiring a 64-bit x86-based processor and a 64-bit OS if running a virtual machine. For assistance and to report problems with the ISO/OVA files contact . firstname.lastname@example.org
The tutorial will use the course material of the Score-P-based tool suite available here.