
in the June 2003

23
June 2003
178 of the June 2003 TOP500
use Myrinet
Myricom's Myrinet® cluster-interconnect technology made another significant
gain in the June 2003 TOP500 list. This list includes 51 Beowulf-style Myrinet
clusters, up from 28 in the November 2002 list. In addition, the list includes
127 HP Superdome/Hyperplex systems. HyperFabric and Hyperplex
are HP brands for Myrinet communication between HP Superdome SMP hosts.
A total of 178 (35.6%) of the June 2003 TOP500 supercomputer sites use
Myrinet technology, up from 140 (28%) in the November 2002 list.
Seventeen of the top 100 systems are Myrinet clusters, led by the 768-node,
1536-processor system at the Forecast System Laboratory, NOAA, which ranks
#11 at 3337 Gflops.
The TOP500 list (www.top500.org),
published twice a year prior to the ISC conference in June and the SC conference
in November, ranks supercomputers worldwide according to their performance
on the LINPACK benchmark. This 21st TOP500 list was published on 23 June
prior to the 18th International Supercomputer Conference (ISC2003)
in Heidelberg, Germany.
Clusters are the fastest growing category of supercomputers in the TOP500.
A cluster first appeared in the TOP500 list in 1997 with the Berkeley NOW
(Network of Workstations), a Myrinet cluster of 100 UltraSPARC-I computers
that ranked #479 in the June 1997 TOP500 list with a LINPACK performance
of 10.14 Gflops. According to the classification used by the TOP500 authors,
there are 149 clusters in the June 2003 list, up from 93 clusters in November
2002, 80 in June 2002 , 43 in November 2001, and 33 in June 2001.
Although nearly all of the TOP500 supercomputers are distributed-memory
systems, the TOP500 authors reserve the cluster classification for systems
in which the interconnect is not proprietary. Thus, a system such as the
IBM SP, which is a cluster architecture that uses IBM-proprietary interconnect,
is classified as an MPP rather than as a cluster. HP HyperFabric is also
classified as a proprietary interconnect.
Here is a rundown on the top 30 Myrinet clusters in the June 2003 TOP500
list. All of these systems use Linux, and most of the newer and higher
ranked systems use Intel® Xeon processors. This list is interesting
for the wide geographic distribution of these clusters, their variety of
purposes, and the diversity of suppliers, including six self-made clusters.
- #11 at 3337 Gflops is a 768-node cluster of dual 2.2GHz Xeon
hosts at the Forecast Systems Laboratory of NOAA. HPTi and Aspen Systems
were the integrators. This newest Myrinet/Linux cluster at FSL, like its
predecessors, is used for weather forecasting.
- #19 at 2207 Gflops is the 512-node "SuperMike" cluster
of 1.8GHz dual Xeon hosts at Louisiana State University (LSU). SuperMike
was named for Louisiana's Governor M. J. "Mike" Foster, Jr.,
who was instrumental in starting an IT initiative to upgrade the state's
computing resources for research and education. This Myrinet/Linux cluster
was supplied by Atipa Technologies.
- #25 at 2004 Gflops is a 300-node Myrinet/Linux cluster of Dell
PowerEdge 2650 (dual 2.4GHz Xeon) servers at the University of Buffalo,
SUNY, Center for Computational Research. This cluster's LINPACK performance
is an unusually high fraction of the Xeon's peak performance due to the
use of BLAS specially coded for the Xeon. The OEM/Integrator for this cluster
was Dell.
- #31 at 1593 Gflops is the latest Titan cluster, a 256-node Myrinet/Linux
cluster of IBM dual 1 GHz Itanium2 hosts at NCSA, part of the Distributed
Terascale Facility. The OEM/Integrator for this cluster was IBM.
- #32 at 1504 Gflops is the Vplant cluster supplied by Dell to
Sandia National Laboratory in Albuquerque. This Myrinet/Linux cluster of
330 dual Xeons is used for visualization, but it also has plenty of compute
performance.
- #51 at 1058 Gflops is the AVIDD cluster at the University of
Indiana, a total of 192 IBM xSeries Linux servers (dual 2.4GHz Xeon) supplied
and integrated by IBM, and interconnected both with Myrinet and with Force10
10Gbit ethernet.
- #52 at 1046 Gflops is a 256-node dual 2.0GHz Xeon cluster at
the Academy of Mathematics and Systems Science in Beijing, China. The Legend
Group was the integrator of this Myrinet/Linux cluster.
- #56 at 1028 Gflops is a 256-node IBM Netfinity (dual 2GHz Xeon)
cluster at the University of Southern California. IBM was the OEM/integrator
of this Myrinet/Linux cluster.
- #58 at 1007 Gflops is the LCRC cluster at Argonne National Laboratory.
This Myrinet/Linux cluster of 361 2.4GHz Xeon processors was supplied by
Linux NetworX.
- #60 at 996.9 Gflops is the Cplant cluster at Sandia National
Laboratory in Albuquerque. This pioneering, self-made, Myrinet/Linux cluster
used 1,800 single-processor Alpha hosts for this TOP500 benchmark. CPlant
first appeared in the TOP500 in the November 1998 list, ranked #97 at 54.24
Gflops achieved with 150 Alphas.
- #79 at 840.5 Gflops is the Grendels 2.4GHz Pentium4 cluster
at Los Alamos National Laboratory, another Linux/Myrinet cluster supplied
by Linux NetworX.
- #81 at 825 Gflops is the HELICS cluster at the University of
Heidelberg, Germany, Interdisciplinary Center for Scientific Computing
(IWR). This cluster of 256 dual 1.4GHz Athlon(TM) hosts (512 processors)
was installed by Megware in March 2002. At #35, HELICS was the top Myrinet
cluster in the June 2002 TOP500 list. All of the clusters listed above
are new or expanded since the TOP500 list of just one year ago.
- #83 and #84 at 798.3 Gflops are twin, 128-node, dual 1GHz Itanium2
clusters at UCSD San Diego Supercomputer Center, two more of the Distributed
Terascale Facility systems (see #31 above). IBM was the OEM and integrator
of these Myrinet/Linux clusters. In comparison with #31, note the near-perfect
scaling (2 times 798.3 Gflops is 1596.6 Gflops, whereas the 256-node NCSA
DTF cluster reached 1593 Gflops).
- #88 at 788.8 Gflops is the self-made Presto III cluster at TITECH
(Tokyo Institute of Technology). This constantly growing Myrinet/Linux
cluster currently has 486 Athlon processors at up to 1.6GHz.
- #89 at 761.9 Gflops is a 128-node cluster of HP zx6000 (dual
900MHz Itanium2) hosts at the Ohio Supercomputer Center. HP was the OEM/integrator
of this Myrinet/Linux cluster, which just barely missed being included
in the November 2002 list.
- #97 at 734.6 Gflops is the MVS1000M cluster at the Joint Supercomputer
Center in Moscow. This self-made Myrinet/Linux cluster of 384 dual 667MHz
Alpha EV-67s is the most powerful computer in Russia.
- #104 at 682.6 Gflops is a Myrinet/Linux cluster of 160 IBM xSeries
(dual 2.4GHz Xeon) servers (128 Myrinet) at Compagnie Generale de Geophysique
(CGG) in the UK.
- #105 at 682.6 Gflops is a Myrinet/Linux cluster of 128 IBM xSeries
(dual 2.4GHz Xeon) servers at Brigham Young University.
- #106 at 682.6 Gflops is yet another Myrinet/Linux cluster of
128 IBM xSeries (dual 2.4GHz Xeon) servers at the South Australian Partnership
for Advanced Computing (SAPAC), Adelaide .
- #107 at 680.3 Gflops is the Demeter cluster of 240 2.6GHz Pentium4
Xeon processors at the American Museum of Natural History. Linux Labs was
the integrator of this Myrinet/Linux cluster.
- #111 at 677.9 Gflops is the original Titan cluster at NCSA.
IBM was the OEM/integrator for this Myrinet/Linux cluster of 160 IBM dual
800MHz Itanium-1 hosts (320 processors). Titan first appeared in the November
2001 TOP500 list ranking #34.
- #119 at 671.3 Gflops is a self-made Myrinet/Linux cluster of
256 2.4GHz Xeon processors at the Fermi National Accelerator Laboratory
in Batavia, Illinois.
- #121 at 654.7 Gflops is a 128-node IBM xSeries (dual 2.0GHz
Xeon) cluster at the Fraunhofer Institut, Ernst-Mach Institute in Freiburg.
- #122 at 654 Gflops is the Magi cluster at the Tsukuba, Japan,
Advanced Computing Center. This Myrinet/Linux cluster of 520 NEC dual 933MHz
Pentium-III hosts (1040 processors) was integrated by NEC in 2001, and,
like its sister system below, runs SCore software developed at RWCP.
- #124 at 629.7 Gflops is a 96-node IBM xSeries (dual 2.8GHz Xeon)
cluster at the Victorian Partnership for Advanced Computing (VPAC), Melbourne.
- #131 at 618.3 Gflops is the SCore IIIe cluster at the Real World
Computing Partnership, Tsukuba, Japan. This self-made Myrinet/Linux cluster
of 512 NEC dual 933MHz Pentium-III hosts (1024 processors) was installed
in 2001, and ranked #36 in the June 2001 TOP500 list.
- #134 at 606.9 Gflops is a Myrinet/Linux cluster of 270 2GHz
Xeon hosts at the University of Oklahoma, Norman. Aspen systems was the
integrator.
- #135 at 594 Gflops is the Platinum Netfinity cluster at NCSA.
This Myrinet/Linux cluster of 512 IBM dual 1GHz Pentium-III hosts (1024
processors) was installed by IBM in 2001, and ranked #30 in the June 2001
TOP500 list.
- #141 at 575 Gflops is a self-made Myrinet/Linux cluster of 128
Appro 1200X dual 2.2GHz Xeon hosts at the Naval Research Laboratory in
Washington DC. This cluster's first appearance in the November 2002 TOP500
was ranked #95.

23 June 2003