There is a Sign-up Schedule for this Event:
htt

p://www.cs.utexas.edu/department/webevent/utcs/events/cgi/show_schedule.cgi

?person=PaoloD''Alberto

Speaker Name/Affiliation: Paulo D''Alberto

Carnegie Mellon University

Date/Time: Monday February 19 2007 No

on - 1:00 p.m.

Location: TAY 3.128

Host: Keshav Pingali

Talk Abstract:
I will introduce some of the techniques and optimizat

ions
developed/used for two examples: All-Pair Shortest Path (APSP) and
Strassen Matrix multiplication (SMM). These techniques improve the perform

ance the code portability and the code maintainability(e.g. ; they simpli

fy the code) and they open new opportunities for aggressive compiler optimi

zations. I show that these two examples (APSP and SMM) build on top of high

ly tuned routines such as GotoBLAS and ATLAS enhancing their application ra

nge and performance. Also I will address issues such as algorithm correct

ness and numerical stability
for APSP and SMM.

This work has bee

n developed in collaboration with Prof.Nicolau (UCI).

Speaker Bio:<

br>Currently Post Doctorate Fellow Electrical and Computer Engineering Ca

rnegie Mellon University. Affiliated to the SPIRAL project responsible for
the design and implementation of techniques for the SW-HW Partitioning of

DFT for low-power and high-performance custom systems.

Ph.D. 2005 C

omputer Science University of California Irvine
Dottorato di Ricerca

2000 Computer Science University of Bologna