SC is the International Conference for
 High Performnance Computing, Networking, Storage and Analysis

SCHEDULE: NOV 13-19, 2010

Combined Iterative and Model-driven Optimization in an Automatic Parallelization Framework

SESSION: Automated Tuning

EVENT TYPE: Paper

TIME: 11:00AM - 11:30AM

SESSION CHAIR: Michela Taufer

AUTHOR(S):Louis-Noel Pouchet, Uday Bondhugula, Cedric Bastoul, Albert Cohen, J. Ramanujam, P. Sadayappan

ROOM:394

ABSTRACT:
Today's multi-core era places significant demands on an optimizing compiler, which must parallelize programs, exploit the memory hierarchy and leverage the ever-increasing SIMD capabilities of modern processors. Existing model-based heuristics for performance optimization used in compilers are severely limited in their ability to identify profitable parallelism/locality trade-offs and usually lead to sub-optimal performance. To address this problem, we distinguish optimizations for which effective model-based heuristics and profitability estimates exist from optimizations that require empirical search to achieve good performance in a portable fashion. We have developed a complete automatic framework in which we focus the empirical search on the set of valid possibilities to perform fusion/code motion, and resort to model-based mechanisms to perform tiling, vectorization and parallelization on the transformed program. We demonstrate the effectiveness of this approach in terms of strong performance improvements on a single target as well as portability of performance to different target architectures.

Chair/Author Details:

Michela Taufer (Chair) - University of Delaware

Louis-Noel Pouchet - Ohio State University

Uday Bondhugula - IBM T.J. Watson Research Center

Cedric Bastoul - University of Paris-Sud

Albert Cohen - INRIA

J. Ramanujam - Louisiana State University

P. Sadayappan - Ohio State University

Add to iCal  Click here to download .ics calendar file

Add to Outlook  Click here to download .vcs calendar file

Add to Google Calendarss  Click here to add event to your Google Calendar

The full paper can be found in the ACM Digital Library and IEEE Computer Society

   Sponsors    IEEE    ACM