UTCS Colloquium: Lars Nyland/NVIDIA Ubiquitous Compute Acceleration (with CS 378) PAR 1 Tuesday February 12 2008 12:30 p.m.

Contact Name: 
Jenna Whitney
Date: 
Feb 12, 2008 12:30pm - 2:00pm

There is a sign-up schedule for this event:
htt

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

br>
Type of Talk: UTCS Colloquium

Speaker/Affiliation: Lars Nyl

and/NVIDIA

Date/Time: Tuesday February 12 2008 12:30 p.m.

Location: PAR 1

Host: Keshav Pingali

Talk Abstract:
In

this talk I will cover three topics: 1) The NVIDIA G80
architecture

2) The CUDA programming language and
3) and recent work on N-Body simu

lation concluding with
examples of similarly accelerated computing for
broader
audiences. The G80 architecture supports both graphics
and
non-graphics computation using an array of custom
processors on a sin

gle chip. The programming model is
neither SIMD nor MIMD but somewhere
in between where
we can exploit the advantages of each. The current

performance part has 128 processors running at 1.3 - 1.9
GHz. With

dual-issue capabilities this places the peak
performance around 500 GFL

OPS. CUDA is the C
programming language with a few extensions for
p

rogramming the G80. These include thread launch
/terminate synchronizat

ion sharing and atomic operations.

In a collaborative effort with

Jan Prins (UNC CS) and Mark
Harris (NVIDIA) we have written an N-Body

simulator using
CUDA that runs on NVIDIA hardware. We achieve a sustain

ed
computational rate over 300 GFLOPS or 16k bodies interacting
at
nearly 50 steps/second. This is substantially faster than a
conventio

nal CPU as the core of the computation relies on
1/sqrt(x) a optimize

d function on the G80 as it is required in
graphics (and physics) for

normalizing vectors. I''ll summarize
with thoughts about the availabil

ity of accelerated computing.

Speaker Bio:
Lars Nyland is a senio

r architect in the ''''compute'''' group at
NVIDIA where he designs

develops and tests architectural
features to support non-traditional us

es of graphics processors.
Prior to joining NVIDIA Lars was an associ

ate professor of
computer science at the Colorado School of Mines in G

olden
Colorado. He ran the Thunder Graphics Lab where demanding

computational applications were coupled with immersive 3D
graphics. B

etween Lars'' PhD and his position in Colorado he
was a member of the

research faculty at UNC Chapel Hill where
he was a member of the high

-performance computing and image-
based rendering groups. Some notable a

chievements were the
development of the DeltaSphere scene digitizer and
its use at
Monticello to provide an immersive experience for visitors

to the
New Orleans Museum of Art''s Jefferson and Napoleon exhibit.
He also spent considerable time studying N-Body algorithms
paralleliz

ing N-Body algorithms for Molecular Dynamics and
parallel programming

languages. Lars earned his PhD at Duke
Univ. in 1991 under the directi

on of John Reif exploring high-
level parallel programming languages.