The
list of supercomputers entered in the Graph500 competition now features
50 competitors, up from nine in its initial release a year ago, said
Sandia National Laboratories researcher Richard Murphy, chair of the
Graph500 steering committee.
New rankings were released Tuesday in Seattle at SC2011, the international conference for high-performance computing.
NNSA/SC
Blue Gene/Q Prototype II has risen to the top spot on the list and is
the first National Nuclear Security Administration winner. Sandia’s
Ultraviolet platform placed 10th using custom software, the Sandia Red
Sky supercomputer dropped from 8th to 13th, and Dingus and Wingus, the
insouciantly named Sandia prototype (Convey-based Field-Programmable
Gate Array, or FPGA) platforms, placed 23rd and 24th, respectively.
The
Graph500 stresses supercomputer performance on “big data” scaling
problems rather than on the purely arithmetic computations measured by
the Linpack Top500 and similar benchmarks. Graph500 machines are tested
for their ability to solve complex problems involving random-appearing
graphs, rather than simply for their speed in solving complex problems.
Such
graph-based problems are found in the medical world, where large
numbers of medical entries must be correlated; in the analysis of social
networks, with their enormous numbers of electronically related
participants; and in international security, where, for example, huge
numbers of containers on ships roaming the world’s ports of call must be
tracked.
“Companies
are interested in doing well on the Graph500 because large-scale data
analytics are an increasingly important problem area and could eclipse
traditional high-performance computing (HPC) in overall importance to
society,” said Murphy, whose committee receives input from 30
international researchers. Changes are implemented by Sandia, the
Georgia Institute of Technology, the University of Illinois at
Urbana-Champaign, Indiana University and others.
Big-data
problems are solved by creating large, complex graphs with vertices
that represent the data points—say, people on Facebook—and edges that
represent relations between the data points—say, friends on Facebook.
These problems stress the ability of computing systems to store and
communicate large amounts of data in irregular, fast-changing
communication patterns, rather than the ability to perform many
arithmetic operations in succession. The Graph500 benchmarks indicate
how well supercomputers handle such complex problems.