Summary: A Unified Framework for Schedule and Storage Optimization
William Thies, Fr´ed´eric Vivien§, Jeffrey Sheldon, and Saman Amarasinghe
Laboratory For Computer Science
Massachusetts Institute of Technology
Cambridge, MA 02139
{thies,jeffshel,saman}@lcs.mit.edu
§ICPS/LSIIT
Universit´e Louis Pasteur
Strasbourg, France
vivien@icps.u-strasbg.fr
Abstract
We present a unified mathematical framework for analyz-
ing the tradeoffs between parallelism and storage allocation
within a parallelizing compiler. Using this framework, we
show how to find a good storage mapping for a given sched-
ule, a good schedule for a given storage mapping, and a
good storage mapping that is valid for all legal schedules.
We consider storage mappings that collapse one dimension
of a multi-dimensional array, and programs that are in a sin-
gle assignment form with a one-dimensional schedule. Our

Source: Amarasinghe, Saman - Computer Science and Artificial Intelligence Laboratory & Department of Electrical Engineering and Computer Science, Massachusetts Institute of Technology (MIT)
Strasbourg, Université de - Département Informatique, ICPS Group

Collections: Computer Technologies and Information Sciences