[BANANA] Berkeley Lab - Scientific Computing Seminar, March 21, 2007

[Parry Husbands]

Date: Wednesday, March 21, 2007
Time: 2:00pm-3:00pm
Location: Building 70-191

Seminar Speaker: Felix Wolf
                  University of Tennessee
                  http://www.cs.utk.edu/~fwolf

Title: Scalable Trace-Based Performance Analysis of Parallel
        Applications

Abstract:

The complex architectures of high-end computing systems present
difficult challenges for performance optimization of scientific
applications. Tools are needed that collect and present relevant
information on application performance so as to enable developers
to easily identify and determine the causes of performance bottlenecks.

Event tracing is a widely used technique for the performance analysis
of parallel applications. Time-stamped events, such as entering a
function or sending a message, are recorded at runtime and analyzed
afterwards. The KOJAK trace analyzer, a joint development of
Forschungszentrum Jlich and the University of Tennessee, can
automatically locate complex performance problems by searching event
traces for patterns of inefficient behavior and quantifying their  
significance.

While KOJAK has proved to provide useful higher-level feedback on
the performance of medium-scale parallel applications, its method of
sequentially analyzing trace data does not scale to applications with
thousands of processes. In the SCALASCA project, we are developing a
scalable version of KOJAK that exploits both distributed memory and
parallel processing capabilities available on the target platform. The
talk will describe the new parallel analyzer architecture and discuss
results for up to 16,384 processes obtained on an IBM Blue Gene/L system.


Sponsor of Seminar: David Bailey