Course on parallelizing compilers, Italy, 29 Jun - 3 Jul

Joseph Edward Hummel <jhummel@broadway.ICS.UCI.EDU>
Sun, 26 Apr 1992 22:14:45 GMT

          From comp.compilers

Related articles
Course on parallelizing compilers, Italy, 29 Jun - 3 Jul jhummel@broadway.ICS.UCI.EDU (Joseph Edward Hummel) (1992-04-26)
| List of all articles for this month |

Newsgroups: comp.compilers
From: Joseph Edward Hummel <jhummel@broadway.ICS.UCI.EDU>
Keywords: parallel, courses
Organization: Compilers Central
Date: Sun, 26 Apr 1992 22:14:45 GMT

Last Call for Participation




                                      The Leonardo Fibonacci Institute
                                  for the Foundations of Computer Science




                                Course on Parallelizing Compilers
                            Trento, Italy, June 29-July 3, 1992




Lecturers:
Dr. U. Banerjee, Intel Corp., Santa Clara
Doctor Banerjee received the Ph.D. in Mathematics from Carnegie-Mellon
University and the Ph.D. in Computer Science
from the University of Illinois. He is currently a Senior Researcher with
the Intel Corporation, Santa Clara. His research interests are in the areas
of dependence analysis and theory of loop transformations.


Prof. A. Nicolau, University of California, Irvine
Professor Nicolau received the Ph.D. in Computer Science from Yale University.
He is currently an Associate Professor
of Computer Science at the University of California, Irvine. His research
interests are in the areas of instruction-level
parallelism, program transformations and parallel architectures.


Prof. D. Padua, University of Illinois, Urbana
Professor Padua received the Ph.D. in Computer Science from the University of
Illinois. He is currently an Associate
Director of CSRD and an Associate Professor in the Department of Computer
Science at the University of Illinois. His research interests are in the
areas of parallelizing compilers, parallel languages, and performance
evaluation.




                                                              Course Outline


For the effective use of today's high-performance machines, excellent code
generation is essential, and the increasing complexity of these machines
demands good parallelizing compilers. The design of such compilers is a
challenging task because it requires integrated solutions not only to
those problems faced by traditional compilers, but also to the new
problems presented by the detection and control of parallelism. Topics to
be covered include the following:


    o Overview of State of the Art Approaches to Parallelizing Compilers:
        Parafrase, Rn/Periscope, Bulldog.


    o Mathematical Background: Relations, Graph Theory, Linear Algebra,
        Number Theory.


    o Dependence Analysis, Anti-Aliasing: Exact and Approximate Methods.


    o Program Transformations: General Unimodular, Iteration Space
        Decomposition, Trace and Percolation Scheduling.


    o Relationship between Compilation Techniques and the Supporting Hardware:
        RISC/Pipelined, Vector, Shared and Distributed Memory,
        VLIW/Superscalar Machines.


    o Future trends.




                                                Course Structure and Applications


The course will be supplemented by guest lectures presented by leading
specialists visiting the Institute. To foster a congenial and productive
atmosphere, attendance is limited to approximately twenty to twenty five
participants. Participants are encouraged to talk about their own
research. A select group of four to six participants will stay on for
additional three weeks and will conduct research in collaboration with the
lecturers.


There are no fees for the course. Ten to fifteen scholarships will be
awarded to outstanding participants, covering local expenses (meals and
lodging). In addition to these scholarships from the Institute, a few
NSF-sponsored scholarships are available for exceptionally qualified
US-based participants. To apply, send a short resume and a cover letter
stating whether you wish to attend only the course or if you would like to
participate in the research experience. If you would like to be considered
for financial support, arrange for two letters of recommendation to be
sent directly, and indicate this in your cover letter. Address all
material to:


              Prof. Alex Nicolau
              Information and Computer Science Department
              University of California, Irvine
              Irvine, CA 92717, U.S.A.


              Phone: (714) 856-4079
              Fax: (714) 856-4056
              Email: nicolau@ics.uci.edu


Applications requesting financial support should be received by April
30th, 1992, and notification of a decision will be sent by May 10th. Other
applications will be considered until May 15th.


Trento is a pleasant historical town in the region of Trentino-Alto Adige,
in the foothills of the Dolomite mountains and next to Lake Garda,
well-connected to Milan, Venice, and Austria.
--
Joe Hummel
ICS Graduate Student, UC Irvine
Internet: jhummel@ics.uci.edu
--


Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.