The Worlds Top Ten Super Computers As Of Nov15
Apple Magic Keyboard, Magic Mouse2, Magic Trackpad2
24 November 2015The History Of Hard Drives
11 December 2015
Here at Geex we love computers, so lets not get started about how much we love Super Computers.
In the last few years (and which we mean 5/6) the power of these SC has grown well beyond what was expected.
They are used for everything from Global Warming, Weather, Nuclear warfare, and Fuel Efficiency simulations to name a few.
The first Super Computer was named Colossus and housed in Britain. It was used to read and crack German messages in World War 2, it could read 5,000 characters a second. Now the Worlds No1. Can read 33,000,000,000,000,000 a second. (that’s a lot of emails)
In 2009 the first Super Computer to top 1 Petaflops/s was called Roadrunner and built by IMB. Now there are 68 systems that can perform over 1 pf/s/
The No.1 spot accounts for 10% of the computing power on the Top 500 list.
In the Top 500 list, HP 179, IBM 110, Cray Inc 71
In fifteen years time it is theorized that the fastest Super Computer will be more powerful than every laptop on the planet today which is a jaw dropping statement.
No. 10 – Stampede – Texas, United States
Pflops/s – 5.168, MAX 8.520
Built By – Dell
Operating System – Linux (CentOS)
Processor Type – Xeon E5-2680 (8Core) + Xeon Phi
Cores – 522,080
Nodes – 6400
Total Memory – 270TB RAM
Disk Storage – 14 PB
Power Usage – 4.5 MW
Number of racks – 182
Approx Weight – 500,000 lbs
No. 9 – Shaheen II – Saudi Arabia Cost $80,000,000
Pflops/s – 5.537, MAX 7.235
Built by – Cray Inc.
Operating System – Linux (CLE)
Processor Type – Xeon E5-2698v3
Cores – 197,568
Nodes – 6174
Total Memory – 790 TB RAM
Disk Storage – 17 PB
Power Usage – 2.8 MW
No. 8 – Hazel Hen – Germany
Pflops/s – 5.640, MAX 7.404
Built by – Cray Inc.
Operating system – Linux (CLE)
Processor Type – Xeon E5-2680v3
Cores – 185,088
Nodes – 7,712
Total Memory – 965 TB RAM
Disk Storage – 11 PB
Disk Drives – 8,300
Power Usage – 7.4 MW
No. 7 – Piz Daint – Switzerland
Pflops/s – 6.271, MAX 7.779
Built By – Cray Inc.
Operating System – Linux (CLE)
Processor Type – Cray Xc30
Can computer in 1 day what a laptop would take 900 years
Nodes – 5,272
cores – 115,000
System Memory – 169 TB RAM
Disk Storage – 2.5 PB
Power usage – 7.7MW
No. 6 – Trinity – United States Cost $174,000,000 (due late 15/2016)
Pflops/s – 8.101, MAX 11.079
Built by – Cray Inc.
Operating System – Linux (CLE)
Processor Type – Xeon E5-2698v3
Cores – 300,000
Disk Space – 82 PB
Used to simulate the safety, security and effectiveness of U.S nuclear deterrent
No. 5 – Mira – United States $50,000,000
Pflops/s – 8,586, MAX 10.066
Built by – IBM
Operating System – Linux (RHEL and CNK)
Processor Type – PowerPC A2
Cores – 786,000
Weight – 100 Tons
No. 4 – K Computer – Japan
Pflops/s – 10.510, MAX 11.280
Built by – Fujitsu
Operating System – Linux
Processor Type – SPARC64 vIIIfx
Cores – 705,024
Cabinets – 864
Nodes – 82,944
Power – 12.6 MW
Annual running cost of $10,000,000
No 3. – Sequoia – United States – Cost $655,000,000
Pflops/s – 17.173 Max 20.133
Built by – IBM
Operating system – Linux (RHEL and CNK)
Processor Type – PowerPC A2
Cores – 1,572,864
Racks – 96
Nodes – 98,304
Memory – 1.6 PB RAM
Power Usage – 7.8 MW
Used for Nuclear weapons simulation.
If everyone on earth used hand held calculators to work on a calculation every second of every day for 320 years, it would take Squoia 1 hour.
No. 2 – Titan – United States – Cost $97,000,000
PFlops/s – 17.59 MAX 27.11
Built By – Cray Inc.
Processor – Opteron 6274 + Tesla K20X
Cores – 560,000
Memory – 710 TB RAM
Disk Space – 40 PB
Nodes – 18,688
Power Usage – 8.2 MW
No. 1 – Tianhe -2 – China Cost $390,000,000 (Available to lease for about $3m a week)
Pflops/s – 33.86, MAX 54.90
Built by – China National University of Defence Technology
Operating System – Linux (Kylin)
Processor – Xeon E5-2692 + Xeon Phi 31s1P
Cores – 3,120,00
Disk Storage – 12.4 PB
Memory – 1 .4 PB RAM
Cabinets – 170
Power Usage – 17.6 MW, With cooling 24 MW
Its worth noting that although this is the no.1 in terms of calculation speed, the ease of use is not good for researchers. It could take a decade to write some of the code needed. The super computers in the US and Japan are far in front of Tianhe-2 in terms of practicality.