MAQAO: une suite d'outils pour l'analyse et l’optimisation des performances / a performance analysis and optimization toolsuite
Andrés S. CHARIF-RUBIAL
Dr. Ingénieur expert, Exascale Computing Research Lab
Abstract : MAQAO (Modular Assembly Quality Analyzer and Optimizer) is a performance analysis and optimization tool suite targeting binary applications (no recompilation). The main goal of MAQAO is to analyse binary codes and provide application developers with synthetic reports in order to help them optimize their code.
The tool mixes both static (code quality evaluation) and dynamic (profiling, characterization) analyses based on the ability to reconstruct low level (basic blocks, instructions, etc.), focusing on high level structures such as functions and loops. Another key feature of MAQAO is its extensibily. Users can easily write their own plugins thanks to an embedded scripting language (Lua). It allows fast prototyping of new MAQAO-tools.
|
Biographie :Andrés S. CHARIF-RUBIAL Dr. Andres S. Charif-Rubial is currently the head of the performance evaluation team at the Exascale
computing research laboratory which is mainly focused on software exascale research and on exascale
application development.
He has been working on the development of the MAQAO toolsuite to better understand the performance issues
of single and multinode HPC applications. His current work aims at defining a consistent performance evaluation
methodology that can deal with the performance evaluation of applications from the core to the multinode level.
He graduated in computer science at the University of Versailles (France) for both his master degree (2008) and his Ph.D. (2012).
During his thesis he strongly focused on coarse grained profiling and memory behavior issues on shared memory systems.
|
|