OPN Talks with …
Alan Gara
IBM Fellow, Blue Gene Chief
Architect and OFC/NFOEC
Keynote Speaker
Courtesy of Alan Gara
As computer technology advances at he speed of light, today’s supercomputer often becomes tomorrow’s personal computer. Many of these advances
would not have been possible without
Alan Gara, who dedicated his career to
being at the forefront of supercomputing. Today, Gara is an IBM Fellow and
the chief architect of the Blue Gene
supercomputer system at IBM’s T.J.
Watson Research Center in New York,
where he is leading the industry in high-performance computing architecture and
design efforts.
Gara will discuss the future of
supercomputers and the role that optical
communications technologies will play
in this field during his plenary session
keynote address at this year’s Optical
Fiber Communication Conference and
Exposition and National Fiber Optic
Engineers Conference (OFC/NFOEC),
6-10 March in Los Angeles.
“Communications fabrics will need to be based on optical
technologies, since there is
a surface-to-volume effect
that requires that large
bandwidth be brought out
of small physical volumes.”
own machines to fit their applications.
For me, computer science began as a tool
to reach my ultimate goal of developing a
cost-effective supercomputer for QCD.
You found that traditional
communications systems
were not enough to power
supercomputer performance.
What is it about optical systems
that makes them more effective?
Communications fabrics will need to
be based on optical technologies, since
there is a surface-to-volume effect that
requires that large bandwidth be brought
out of small physical volumes. Higher
computing densities result in less overall
communications bandwidth and hence
more cost-optimized solutions.
in the software required to program and
control massively parallel systems, and
in the use of computation to advance our
understanding of important biological
processes such as protein folding.
The Blue Gene/L machine was
designed and built in collaboration with
the Department of Energy’s NNSA/
Lawrence Livermore National Laboratory in California, and the LLNL system
has a peak speed of 596 Teraflops. Blue
Gene systems occupy the #4 (LLNL Blue
Gene/L) and #5 (Argonne Blue Gene/P)
positions in the Top 500 supercomputer
list announced in November 2008.
IBM and its collaborators are currently exploring a growing list of applications
including hydrodynamics, quantum
chemistry, molecular dynamics, climate
modeling and financial modeling.
What inspired you to go into
computer science?
I entered computer science research from
the not-uncommon direction of physics
simulation. There have been a number
of Grand Challenge problems (goals set
in the late-1980s for high-performance
computing and communications), such
as quantum chromodynamics (QCD),
that require nearly unlimited computational resources. This need for inexpensive large-scale computing drove me and
many other physicists to develop their
How do you think the advances
in optical technologies will affect
the greater commercial market?
The challenges of supercomputing will
likely be mirrored in the broader commercial market at a later time. It will
be mostly a cost issue and the advances
driven by supercomputing will likely
carry over naturally into this broader
commercial market.
What exactly is the Blue
Gene system?
Blue Gene is an IBM research project
dedicated to exploring the frontiers in
supercomputing: in computer architecture,
What does the future of
supercomputing look like?
The challenges for supercomputers are
enormous. There are discontinuous
technology impacts—both positive and
18 | OPN Optics & Photonics News
www.osa-opn.org
