Australia's Victorian Life Sciences Computation Initiative hosts the fastest life science supercomputer in the world, according to the most recent version of the Top500 list released earlier this month.
The 65,536-core IBM BlueGene/Q system, known as "Avoca," clocks in at 690.2 teraflops and holds the No. 33 position on the list of the world's fastest supercomputers, a slight drop from the No. 31 spot on the June version of the twice-yearly ranking.
The four other life science systems on the list fell considerably since June, with the second fastest, the 16,128-core, 100.6-teraflop Shirokane2 system at the University of Tokyo's Human Genome Center, dropping to the No. 301 spot from No. 183 just six months ago.
The remaining life science systems on the list include a 97.1-teraflop, 18,176-core HP cluster at Pacific Northwest National Laboratory’s Environmental Molecular Sciences Laboratory, which fell to No. 313 in the current list from No. 193 in June; a 9,498-core, 89.8 teraflop Rackable cluster at the Polish Academy of Science's Institute of Bioorganic Chemistry, which fell to No. 375 from No. 233; and an 82.9-teraflop, 5,616-core HP system at Japan's National Institute of Genetics, which dropped to No. 432 from No. 280 in June (see Table 1, below, for details of the life science systems on the list).
The fastest computer on the list is Titan, a Cray XK7 system installed at Oak Ridge National Laboratory, which achieved 17.59 petaflops on the Linpack benchmark used to assess the performance of the Top500 systems.
Titan, an upgrade of ORNLs' previous Jaguar system, has 560,640 processors, including 261,632 Nvidia K20x accelerator cores.
Titan's reliance on Nvidia GPUs underscores the move toward accelerator and co-processor technology in high-performance computing. A total of 62 systems on the Top500 list are now using some sort of accelerator technology, with the majority — 50 — opting for some type of Nvidia chip. Six months ago, 58 systems on the Top500 list used accelerators or co-processors.
Furthermore, systems with multi-core processors continue to dominate the list, with 85 percent of systems in the current list using processors with six or more cores and 46 percent using processors with eight or more cores.
Intel continues to provide the most processors for high-performance computing, with 380 systems on the current list running Intel chips, up slightly from 384 systems in the June list. Following Intel is the AMD Opteron family, with 60 systems on the current ranking — the same as six months ago.
IBM continues to manufacture the most systems in the Top500 list, thought its lead over HP slipped slightly in the current ranking. IBM claims 193 systems in the current list, compared to 213 in June, while HP now claims 146 Top500 systems, compared to 138 in June (see Table 2, below, for details of manufacturer rankings).
Clusters are still the most popular architecture for supercomputers in the Top500 list, with 411 systems in the current list. Massively parallel processing systems comprise the balance of the systems.
Linux continues to be the top operating system for Top500 machines, with 469 computers in the current list, a slight bump from 462 in the June ranking. Unix systems remained the next most popular, but slid a bit to 20 systems from 24 six months ago.
The complete Top500 ranking is available here.