TAU Training at ORNL
This workshop will cover performance evaluation of applications on OLCF’s HPC platforms. This workshop will focus on performance data collection, analysis, and performance optimization. After describing and demonstrating how performance data (both profile and trace data) can be collected in a straightforward manner using TAU’s (Tuning and Analysis Utilities) automated instrumentation, the workshop will cover how to analyze the performance data collected and drill down to find performance bottlenecks and determine their causes.
The workshop will include some sample codes that illustrate the different instrumentation and measurement choices available to the users. Topics will cover generating performance profiles and traces with GPUs using CUDA, OpenCL, and CUPTI. It will include memory utilIzation and headroom, I/O and hardware performance counters data using PAPI.
Instructor: Sameer Shende
April 15, 2014:
9:00am – 9:45am: Setting up accounts, setting up environment to run workshop examples.
9:45am – 10:00am: Short break
10:00am – 12:00pm: Introduction to TAU- Instrumentation options, demonstration
12:00 -1:30pm Hands-on: MPI, ParaProf
1:30pm – 3:30pm: Instrumentation and Measurement options in TAU – Compiler-based instrumentation, CUDA, Score-P, demonstration
3:30pm – 3:45pm: Break
3:45pm – 5pm: Hands-on session with TAU – Profiling workshop examples and user applications
April 16, 2014:
9am – 10:30am: Memory and I/O instrumentation, PAPI hardware performance counters – Demonstration of debugging features
10:30 am – break
10:45am – 12:45pm:Analysis tools: PerfExplorer, TAUdb
12:45am – 2:45pm Hands-on
2:45pm – 3:15pm: Break
3:15pm – 5pm: Trace visualization: Vampir and Jumpshot and hands-on.