Two new life science computers have debuted in the most recent Top500 ranking of the world's fastest supercomputers, joining three other life science machines that surpassed the 24.7-teraflop benchmark for inclusion on the current version of the twice-yearly list.
One of the new systems is an 8,640-core system at the Georgia Institute of Technology's Center for the Study of Systems Biology that was build by Penguin Computing. The computer, ranked at No. 98 on the Top500 list, clocked in at 53.1 teraflops. Jeffrey Skolnick, who leads the systems biology center, is using the system for large-scale computer simulations of proteins and cell models.
The other new entry is a 35.8-teraflop, 4,000-core Dell system called "Gladiator" that was installed on the Howard Hughes Medical Institute's Janelia Farm Campus earlier this year to support a range of life science research activities.
A 97.1-teraflop, 18,176-core HP cluster at Pacific Northwest National Laboratory’s Environmental Molecular Sciences Laboratory remains the fastest life science machine on the list and holds the No. 57 spot — down from No. 48 in the last version of the ranking, which was released in November (BioInform 12/4/2009).
Two other life science systems remain from the November list: a 54-teraflop, 5,760-core Sun Microsystems blade system at the University of Tokyo's Human Genome Center that now holds the No. 96 spot, down from No. 84 six months ago; and a 30.1-teraflop Dell system at Arizona State University and the Translational Genomics Research Institute, which fell to No. 332 from No. 222 in the most recent ranking (see Table 1, below, for details of the life science systems on the list).
A 24-teraflop, 5,712-core, HP cluster at health informatics firm Cerner that debuted on the November list at the No. 375 spot did not make the 24.7-teraflop cutoff for the current version.
The fastest machine in the world remains the "Jaguar" Cray XT5 supercomputer at the Department of Energy’s Oak Ridge Leadership Computing Facility, which has a maximum speed of 1.75 petaflops and a theoretical peak capability of 2.3 petaflops.
A new system, China's "Nebulae," debuted at the No. 2 spot. The machine, housed at the National Supercomputing Centre in Shenzhen, China, is a Dawning TC3600 Blade system with Intel X5650 processors and NVidia Tesla C2050 GPUs. It achieved 1.271 petaflops on the Linpack benchmark, which puts it behind Jaguar on the Top500 list, but its GPU accelerators give it an edge over Jaguar in terms of theoretical peak capability, which is the highest ever on the Top500 list at nearly 3 petaflops.
IBM and HP remain the dominant supercomputing vendors. HP lost a slight lead that it held over IBM in the November list and now has 185 out of the top 500 systems, while IBM claims 198 systems. HP had 210 installed systems on the list six months ago, while IBM had 186.
Cray and Dell follow, with 21 and 17 systems on the current list, respectively (see Table 2, below, for details on manufacturer rankings).
Clusters remain the most popular architecture for systems in the Top500 list, with 424 cluster-based systems in the current list compared to 417 six months ago (see Table 3, below, for details).
Intel processors are dominant, with 406 machines on the current list using Intel chips as opposed to 402 six months ago. AMD claims 49 systems, up from 42 six months ago; while there are 42 IBM Power processors on the current list, down from 52 in November (see Table 4, below, for details).
In terms of operating systems, Linux now claims 91 percent of all machines on the Top500 list, with 455 systems compared to 446 in the November ranking. Unix systems were the next most popular, with 22 systems on the current list, followed by "mixed" operating systems with 17.
The complete Top500 ranking is available here.