MIT Department of Electrical Engineering & Computer Science
Automatic Compilation for Multiprocessors
Saman Amarasinghe
Stanford University
Thursday, April 11, 1996
3:00 PM (2:45 refreshments)
Room NE43-518
EECS Special Seminar
Abstract
We have developed SUIF, a parallelizing and optimizing compiler that
uses whole program analysis and aggressive optimizations to obtain
performance improvements for significantly more scientific
applications than that was previously possible. Many of these crucial
analysis and optimization techniques were developed using a common
mathematical framework based on integer programming. In this talk, I
will describe how this framework was used effectively to perform
analyses and optimizations such as array data-flow analysis,
communication code-generation and data transformation optimizations.
I will show the performance improvements we have obtained for several
benchmark suites on shared memory multiprocessors that demonstrates
the effectiveness of these novel techniques.
Host: Prof. Anant Agarwal
URL of this page:
http://www-eecs.mit.edu/AY95-96/events/46.html
Created: Apr 4, 1996
|
Modified: Jun 25, 1997
This announcement is from the MIT EECS 1995-96 archive.
|
Current events
To MIT EECS home page
|
Your comments
and inquiries are welcome.
