The TOP10 shows five new systems and one substantially upgraded system with five of these changes placing at the top five positions.
The new No. 1 system is a significantly enlarged version of the previous No. 1 - the DOE's IBM BlueGene/L system, installed at DOE's Lawrence Livermore National Laboratory (LLNL) with a Linpack performance of now 478.2 TFlop/s compared to 280.6 TFlop/s six months ago before its upgrade.
At No. 2 is a brand-new first installation of a newer version of the same type of IBM system. It is a BlueGene/P system installed in Germany at the Forschungszentrum Juelich (FZJ) and it achieved 167.3 TFlop/s.
The No. 3 system is not only new but the first system for a new center, the New Mexico Computing Applications Center (NMCAC) in Rio Rancho, NM. It is built by SGI and based on the Altix ICE 8200 model. It was measured at 126.9 TFlop/s.
For the first time ever, India placed a system in the top10. The Computational Research Laboratories, a wholly owned subsidiary of Tata Sons Ltd. in Pune, India, installed a Hewlett-Packard Cluster Platform 3000 BL460c system. They integrated this system with their own innovative routing technology and achieved a performance of 117.9 TFlop/s.
This is the first system outside the US, Europe, or Japan to ever enter the TOP5.
The No.5 system is also a new Hewlett-Packard Cluster Platform 3000 BL460c system and installed at a Swedish government agency. It was measured at 102.8 TFlop/s.
The last new system in the TOP10 - at No. 9 - is a Cray XT4 system installed at the National Energy Research Scientific Computing Center (NERSC) at DOE's Lawrence Berkeley National Laboratory and was ranked based on a Linpack performance of 85.4 TFlop/s.
General highlights from the Top 500 since the last edition
The entry level to the list moved up to the 5.9 TFlop/s mark on the Linpack benchmark, compared to 4.0 TFlop/s six months ago.
The last system on the list would have been listed at position 255 in the previous TOP500 just six months ago. This turnover rate is above average.
Total combined performance has grown to 6.97 PFlop/s, compared to 4.92 PFlop/s six months ago and 3.54 PFlop/s one year ago.
The entry point for the top 100 increased in six months from 9.29 TFlop/s to 12.97 TFlop/s.
A total of 354 systems (70.8 percent) are now using Intel processors. This is up from six months ago (289 systems, 57.8 percent) and a represents the largest share for Intel chips in the TOP500 ever.
The AMD Opteron family, which passed the IBM Power processors a year ago, remained the second most common processor family with 78 systems (15.6 percent), down from 105 systems (21 percent) six months ago. 61 systems (12.2 percent) use IBM Power processors, down from 85 systems (17 percent) six months ago.
Multi core processors are the dominant chip architecture. The most impressive growth showed the number of systems using the Intel Clovertown quad core chips, which grew in six months from 19 to 102 systems.
The majority of remaining systems uses dual core processors.
406 systems are labeled as clusters, making this the most common architecture in the TOP500 with a stable share of 81.2 percent.
Gigabit Ethernet is still the most-used internal system interconnect technology (270 systems), due to its widespread use at industrial customers, followed by InfiniBand technology with 121 systems.
For quite some time, IBM and Hewlett-Packard sell the bulk of systems at all performance levels of the TOP500.
IBM regained a solid lead this time with 232 systems (46.4 percent) over HP with 166 systems (33.2 percent). IBM, which had been ahead of HP since June 2004, lost the lead in the number of systems six months ago with 38.4 percent (down from 47.2 percent one year ago), compared to HP with 40.6 percent (up from 31.6 percent one year ago).
IBM remains the clear leader in the TOP500 list in performance with 45 percent of installed total performance (up from 41.9), compared to HP with 23.9 percent (down from 24.5).
In the system category, again no other manufacturer could break the 5 percent barrier, but Dell and SGI got very close with 4.8 percent and 4.4 percent.
In the performance category, the manufacturers with more than 5 percent are: Cray (7.4 percent of performance), SGI (7.3 percent), and Dell (7 percent of performance), each of which benefits from large systems in the TOP100.
IBM (141) and HP (140) sold together 281 out of 287 systems at commercial and industrial customers and have had this important market segment clearly cornered for some time now.
Highlights from the Top 50
The U.S. is clearly the leading consumer of HPC systems with 284 of the 500 systems. The European share (149 systems - up from 127) is still rising and is again larger then the Asian share (58 - down from 72 systems).
Dominant countries in Asia are Japan with 20 systems (down from 23), Taiwan with 11 (up from 10), China with 10 systems (down from 13), and India with 9 systems (up from eight).
In Europe, UK has established itself as the No. 1 with 48 systems (43 six months ago). Germany has to live with the No. 2 spot with 31 systems (24 six months ago).
The entry level into the TOP50 is at 19.4 TF/s
The U.S. has about the higher percentage of systems (64 percent) in the TOP50 than in the TOP500.
The dominant architectures are custom build massively parallel systems MPPs with 52 percent ahead of commodity clusters with 44 percent.
IBM leads the TOP50 with 34% of systems and 43 percent of performance.
No 2 is DELL with 16 percent of systems and 10.5 percent of performance.
Cray is third with 12 percent of systems and 13.8 percent of performance closely followed by SGI with 12 percent of systems and 13.4 percent of performance.
HP absent from the TOP50 six month ago has now 10 percent of systems and 10.7 percent of performance.
52 percent of systems are installed at research labs and 34 percent at universities.
There is no system using Gigabit Ethernet in the TOP50.
IBM's BlueGene is the most used system family with 10 systems (20 percent).
Intel processors are used in 48 percent of systems ahead of IBM's Power processors in 30 percent and AMD in 10 percent.
The average concurrency level is 15,690 cores per system up from 11,300 six month ago.
The average age of a TOP50 system is only about 1 year and 6 month. 70% have been installed or upgraded this year and 20 percent last year.
