A Chinese system called Nebulae, build from a Dawning TC3600 Blade system with Intel X5650 processors and NVidia Tesla C2050 GPUs is now the fastest in theoretical peak performance at 2.98 PFlop/s and No. 2 with a Linpack performance of 1.271 PFlop/s. This is the highest rank a Chinese system ever achieved. There are now 2 Chinese systems in the TOP10 and 24 in the TOP500 overall.
China keeps increasing its number of systems to 24 and is now tied with Germany (steadily declining) for spot No. 4 after the USA, UK and France.
China also climbed with respect to overall installed performance and is now holding for the first time the No. 2 spot behind the USA and ahead of Germany.
The Jaguar system at Oak Ridge National Laboratory managed to hold the No. 1 spot with 1.75 PFlop/s Linpack performance even as it’s peak performance is lower than the Chinese Nebulae system.
The most powerful system in Europe is an IBM BlueGene/P system at the German Forschungszentrum Juelich (FZJ) which dropped to No. 5.
Intel dominates the high-end processor market 81.6 percent of all systems and over 90 percent of quad-core based systems.
The Intel Core i7 (Nehalem-EP) processors increased their presence in the list with 186 systems compared with 95 in the last list.
Quad-core processors are used in 85 percent of the systems and 5 percent use already processors with six or more cores.
Other notable systems are:IBM regains the lead in market share by total systems from Hewlett-Packard, IBM also stays ahead by overall installed performance.
The Tianhe-1 system at No. 7, which is a hybrid design with Intel Xeon processors and AMD GPUs. The TH-1 uses AMD GPUs as accelerators. Each node consists of two AMD GPUs attached to two Intel Xeon processors.
The Cray’s XT system series remains very popular for big research customers with 10 systems in the TOP50 (20 percent).
Power consumption of supercomputers
TOP500 now tracks actual power consumption of supercomputers in consistent fashion.
Only 22 systems on the list are confirmed to use more than 1 MWatt of power.
The No. 1 system Jaguar reports the highest total power consumption of 6.95 MWatt.
Average Power consumption of a TOP500 system is 397 kWatt and average power efficiency is 195 Mflops/Watt (up from 150 Mflops/Watt one year ago).
Average Power consumption of a TOP10 system is 2.89 MWatt (up from 2.45 MWatt a year ago) and average power efficiency is 300 Mflops/Watt up from 280 Mflops/Watt one year ago.
Most energy efficient supercomputers are based on
QPace Clusters based on IBM PowerXCell 8i processor blades in Germany (up to 774 Mflop/Watt)
IBM QS22 Cell processor blades (up to 458 Mflop/Watt),
Highlights from the Top 10:
The TOP10 features only one new system and one upgrade
The new Chinese Nebulae system at #2 is only the third system ever to break the Petaflop barrier.
The upgraded Pleiades system installed at NASA Ames Research Center jumped one competitor and is now ranked #6 with 772.7 TFlop/s Linpack performance.
There is second Chinese system in the TOP10: The Tianhe-1 system installed at the National Super Computer Center in Tianjin, China is now #7.
In the TOP10 only the No. 2, 5, and 7 systems are installed outside the U.S. – in this case in China, Germany, and China.
General highlights from the Top 500 since the last edition:
Quad-core processor based systems dominate the TOP500 as 425 systems are using them. 48 systems are still using dual-core processors but already 25 systems are using processors with 6 or more cores.
The entry level to the list moved up to the 24.7 TFlop/s mark on the Linpack benchmark, compared to 20 TFlop/s six months ago.
The last system on the newest list was listed at position 357 in the previous TOP500 just six months ago. This turnover rate is far below average and might reflect the economic situation as well as an upcoming new product cycle in the HPC market.
Total combined performance of all 500 systems has grown to 32.4 PFlop/s, compared to 27.6 PFlop/s six months ago and 22.6 PFlop/s one year ago.
The entry point for the top 100 increased in six months from 47.72 TFlop/s to 52.84 TFlop/s.
The average concurrency level in the TOP500 is 10,267 cores per system up from 9,174 six month ago and 8,210 one year ago.
A total of 408 systems (81.6 percent) are now using Intel processors. This is slightly up from six months ago (402 systems, 80.4 percent). Intel continues to provide the processors for the largest share of TOP500 systems.
They are now followed by the AMD Opteron family with 47 systems (9.4 percent), up from 42.
The share of IBM Power processors is slowly declining with now 42 systems (8.4 percent), down from 52.
424 systems are labeled as clusters, making this the most common architecture in the TOP500 with a stable share of 85 percent.
Gigabit Ethernet is still the most-used internal system interconnect technology (244 systems), due to its widespread use at industrial customers, followed by InfiniBand technology with 205 systems.
However, Infiniband based system account for twice as much performance (15.8 PFlop/s) than Gigabit Ethernet ones (7.8 PFlop/s).
IBM and Hewlett-Packard continue to sell the bulk of systems at all performance levels of the TOP500.
HP lost its narrow lead in systems and has now 185 systems (37 percent) compared to IBM with 198 systems (39.8 percent). HP had 210 systems (42 percent) six months ago, compared to IBM with 186 systems (37.2 percent).
IBM remains the clear leader in the TOP500 list in performance with 33.6 percent of installed total performance (down from 35.1 percent), compared to HP with 20.4 percent (down from 23 percent).
In the system category, Cray, SGI, and Dell follow with 4.2 percent, 3.4 percent and 3.4 percent respectively.
In the performance category, the manufacturers with more than 5 percent are: Cray (14.8 percent of performance) and SGI (6.6 percent), each of which benefits from large systems in the TOP10.
HP (167) and IBM (128) sold together 295 out of 302 systems at commercial and industrial customers and have had this important market segment clearly cornered for some time now.
The U.S. is clearly the leading consumer of HPC systems with 282 of the 500 systems (up from 277). The European share (144 systems – down from 152) is still substantially larger then the Asian share (57 systems – up from 51).
Dominant countries in Asia are China with 24 systems (up from 21), Japan with 18 systems (up from 16), and India with 5 systems (up from 3).
In Europe, UK remains the No. 1 with 38 systems (45 six months ago). France passed Germany and has now 29 (up from 26). Germany is still now the No. 3 spot with 24 systems (27 six months ago).
Highlights from the Top 50:
The entry level into the TOP50 is at 102.8 TFlop/s
The U.S. has a lower percentage of systems (40 percent) in the TOP50 than in the TOP500 (56.4 percent).
The dominant architectures are custom-built massively parallel systems MPPs with 56 percent ahead of commodity clusters with 44 percent.
IBM leads the TOP50 with 24 percent of systems and 27 percent of performance.
No. 2 is Cray with a stable share of 20 percent of systems and 27 percent of performance.
SGI is third with 16 percent of systems and 11.2 percent of performance.
HP has 4 percent of systems and 3.4 percent of performance.
60 percent of systems are installed at research labs and 28 percent at universities.
There is only a single system using Gigabit Ethernet in the TOP50.
Cray’s XT is the most-used system family with 10 systems (20 percent), followed by SGI Altix with 8 systems (16 percent).
Intel processors are now used in 48 percent of systems, ahead of AMD processors in 26 percent and IBM’s Power processors in 22 percent.
The average concurrency level is 49,080 cores per system – up from 44,338 cores per system six month ago and 40,871 one year ago.
All changes are from November 2009 to June 2010.
The TOP500 list is compiled by Hans Meuer of the University of Mannheim, Germany; Erich Strohmaier and Horst Simon of NERSC/Lawrence Berkeley National Laboratory; and Jack Dongarra of the University of Tennessee, Knoxville.