Highlights - November 2012

  • Titan, a Cray XK7 system installed at the Oak Ridge National Laboratory by the Department of Energy (DOE) is the No. 1 system on the TOP500. It achieved 17.59 Petaflop/s on the Linpack benchmark using 261,632 of its NVIDIA K20x accelerator cores. Titan is one of the most energy efficient systems on the list consuming a total of 8.21 MW and delivering 2,143 MFlops/W.

  • Sequoia, the IBM BlueGene/Q system installed at the National Nuclear Security Administration and part of the Advanced Simulation and Computing Program (ASC) is now the No. 2 system on the TOP500. It was first delivered to the Lawrence Livermore National Laboratory in 2011and was full deployed with an impressive 16.32 Petaflop/s on the Linpack benchmark using 1,572,864 cores just six month ago. Sequoia is also one of the most energy efficient systems on the list consuming a total of 7.84 MW and delivering 2,031.6 MFlops/W.

  • Fujitsu’s “K Computer” installed at the RIKEN Advanced Institute for Computational Science (AICS) in Kobe, Japan, is now the No. 3 system on the TOP500 list with10.51 Pflop/s on the Linpack benchmark using 705,024 SPARC64 processing cores.

  • A second BlueGene/Q system (Mira) installed at Argonne National Laboratory is at No. 4 with 8.15 Petaflop/s on the Linpack benchmark using 786,432 cores.

  • A third BlueGene/Q system (JUQEEN) at No. 5 is installed at the Forschungszentrum Juelich in Germany and is listed with 4.14 PFlop/s as most powerful system in Europe.

  • At No. 6 is the second system in Europe , the SuperMUC, an IBM iDataplex system with Intel Sandybridge installed at Leibniz Rechenzentrum in Germany.

  • At No. 7 is the only truly new system in the TOP10 called “Stampede”. It is a Dell PowerEdge C8220 system installed at the Texas Advanced Computing Center of the University of Texas. It uses the brand new intel Xeon Phi processors (previously know as MIC) to achieve its 2.6 PFlop/s.

  • The Chinese Tianhe-1A system, the No. 1 on the TOP500 in November 2010 is now the No. 8 with 2.57 Pflop/s Linpack performance.

  • No. 9 is taken by an IBM BlueGene/Q system at CINECA in Italy.

  • The TOP10 is rounded out by the partially assembled Power7 based IBM system for DARPA.

  • There are 23 Petaflop/s systems in the TOP500 List up from 20 six month ago.

  • The new No 1 system Titan and the No. 8 systems are using NVIDIA GPUs to accelerate computation and the No. 7 system is accelerated by Intel Xeon Phi processors.  A total of 62 systems on the list are using Accelerator/Co-Processor technology, 7 of these are based on the Intel Xeon Phi.

  • The number of systems installed in China has now stabilized at 72 with 68 and 74 on the last two lists.  China has occupied the No. 2 position as a user of HPC, ahead of Japan, UK, France, and Germany. Japan however holds the No. 2 position in the performance share ahead of China.

  • Intel continues to provide the processors for the largest share (75.8 percent) of TOP500 systems.

  • Already 84.4 percent of the systems use processors with six or more cores and 46 percent with eight or more cores.

  • 62 systems use accelerators or co-processors (up from 58 six month ago), 50 of these use NVIDIA chips, two use Cell processors, three use ATI Radeon, and already seven systems with Intel MIC technology (Xeon Phi).

  • IBM’s Bluegene/Q is now the most popular system in the TOP10 with 4 entries including the No. 2, 4, 5, and No. 9.

General highlights from the TOP500 since the last edition:

  • Already 84.4 percent of the systems use processors with six or more cores.

  • Already 46 percent of the systems use processors with eight or more cores.

  • The entry level to the list moved up to the 76.5 Tflop/s mark on the Linpack benchmark, compared to 60.8 Tflop/s six months ago.

  • The last system on the newest list was listed at position 336 in the previous TOP500 just six months ago.

  • Total combined performance of all 500 systems has grown to 162 Pflop/s, compared to 123 Pflop/s six months ago and 74 Pflop/s one year ago.

  • The entry point for the TOP100 increased in six months from 172.7 Tflop/s to 243.9 Tflop/s.

  • The average concurrency level in the TOP500 is 29,796 cores per system, up from 26,327 six months ago and 18,383 one year ago.

  • A total of 379 systems (75.8 percent) are now using Intel processors, about the same as in the in the previous list 6 months ago.

  • Intel is followed by the AMD Opteron family with 61 systems (12.2 percent), down from 62 (12 percent) in the previous list 6 months ago.

  • The share of IBM Power processors is at 53 systems (10.6 percent).

  • 62 systems use accelerators or co-processors (up from 58 six month ago), 50 of these use NVIDIA chips, two use Cell processors, three use ATI Radeon, and already seven systems with Intel MIC technology (Xeon Phi).

  •  
  • InfiniBand technology is now with 225 systems, up from 209 systems the most-used internal system interconnect technology. Gigabit Ethernet is fell back to 188 systems, down from 207 systems.

  • InfiniBand-based systems also account for more than twice as much performance (52.7 Pflop/s) than Gigabit Ethernet ones (20.3 Pflop/s).

 

  • IBM and Hewlett-Packard continue to sell the bulk of the systems at all performance levels of the TOP500.

  • IBM kept its lead in systems and has now 193 systems (38.6 percent) compared to HP with 146 systems (29.2 percent). HP is slightly up from 141 systems (28.2 percent) six months ago, compared to IBM with 215 systems (43 percent).

  • IBM remains the clear leader in the TOP500 list in performance and has a considerably lead with a share 41 percent of installed total performance (down from 49.1 percent).

  • Cray increased its share in performance substantially thanks to Titan’s impressive No. 1 performance and now has 17.4 percent up from 8.8 percent.

  • HP is now only third even so it increased its share to 11.4 percent up from 10.5 percent. 

  • Fujitsu follows closely in the 4th spot due to the impressive performance of the No. 3 K Computer with 8.5 percent down from 10.1 percent.  

  • In the system category, Cray, Appro, SGI, and Bull follow with 6.2 percent, 4.8 percent, 3.8 percent, and 3.6 percent respectively.

 

  • The U.S. is clearly the leading consumer of HPC systems with 250 of the 500 systems (252 last time). The European share (105 systems – 106 last time) is still lower than the Asian share (124 systems – 122 last time).

 

  • Dominant countries in Asia are China with 72 systems (up from 68) and Japan with 31 systems (down from 35).

  • In Europe, UK, France, and Germany, are almost equal with 24, 21, and 19 respectively.