To exploit current and future systems, developers need to understand the
interplay between on-node performance, distribution of application data
among processes, and the application's intrinsic communication patterns.
However, current tools do a poor job in presenting the information required
to analyze this interplay. This talk will present a novel approach that
explicitly addresses this problem. We identify the three main domains most
familiar to the user - the application data or physical domain, the hardware
domain, and the communication domain - and provide mappings for data
gathered in each of the domains to the other two. This not only allows us to
present the data in domains that are more intuitive to the user, but also
enables us to directly correlate metrics from multiple domains using data
analysis techniques such as feature detection. This opens the door for a new
generation of tools that provide performance data in a more intuitive manner
to the application developer.
More Intuitive Performance Analysis
21.09.2011
Ημερομηνία : 21.09.2011
Ώρα : 16:00 - 17:30
Μέρος : Aίθουσα Συναντήσεων Γ100 ΙΤΕ, Ηράκλειο, Κρήτη
Φιλοξενείται από : Μπίλας Άγγελος
Martin 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 Universitaet
Muenchen (Munich, Germany). He also holds a Master of Science in Computer
Science from the University of Illinois at Urbana Champaign. After
completing his graduate studies and a postdoctoral appointment in Munich, he
worked for two years as a Research Associate at Cornell University, before
joining LLNL in 2004. Martin's research interests include parallel and
distributed architectures and applications; performance monitoring, modeling
and analysis; memory system optimization; parallel programming paradigms;
tool support for parallel programming; power efficiency for parallel
systems; optimizing parallel and distributed I/O; and fault tolerance at the
application and system level. In his position at LLNL he especially focuses
on the issue of scalability for parallel applications, code correctness
tools, and parallel performance analyzer as well as scalable tool
infrastructures to support these efforts. Martin is a member of LLNL's ASC
CSSE ADEPT (Application Development Environment and Performance Team) and he
works closely with colleagues in CASC's Computer Science Group (CSG) and in
the Development Environment Group (DEG). He is also 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".