UCRL-PRES-145618-1 Lawrence Livermore National Laboratory , P. O. Box 808, Livermore, CA 94551
August 13, 2002
Welcome to Blue Gene/L
Presented to
Blue Gene/L ConferenceLake Tahoe, CA
David A. NowakDeputy Associate Director for Defense and Nuclear Technologies
Program Leader for Advanced Simulation and ComputingLawrence Livermore National Laboratory
Expanding Our High-End Computing Universe
UCRL-PRES-145618-2
One way of looking at Blue Gene/Lis to look at the physics of scale
10–8 10–4 0.01 110–10 10–6
Distance (m)
Nuclear Physics
Full System
Atomic Physics
Molecular Level
Turbulence
Continuum Models
BBlluueeGGeennee//LL
ASCI
Purple
ASCI
Purple
Blue Gene/L
UCRL-PRES-145618-3
Blue Gene/L differs from current leading edge simulation platforms in some important areas
Blue Gene/L
UCRL-PRES-145618-4
Important points about Blue Gene/L
z Embedded technology promises to be an efficient path toward building massively parallel computers optimized at the system level
z Cost / performance is ~ 20 to 50 times better than standard methods to get to teraflops
z Low Power is critical to achieving a dense, simple, inexpensive packaging solution
z Blue Gene/L will have a scientific reach far beyond existing limits for a large class of important scientific problems
z Blue Gene/L will give insight into possible future product directions
z Blue Gene/L hardware will be flexible
z A mature, sophisticated software environment needs to be developed to really determine the reach (both scientific and commercial) of this architecture
Blue Gene/L