Tianhe-2: Difference between revisions
FrancisWang (talk | contribs) No edit summary |
FrancisWang (talk | contribs) No edit summary |
||
| Line 39: | Line 39: | ||
Tianhe-2 consists of 16,000 compute nodes, each comprising two Intel [[Ivy Bridge (microarchitecture)|Ivy Bridge]] Xeon processors and three [[Intel MIC#Xeon Phi|Xeon Phi]] chips for a total of 3,120,000 cores.<ref name="TOP500June2013"/> This represents the world's biggest installation of Ivy Bridge and Xeon Phi chips. Each of the 16,000 nodes possess 88 gigabytes of memory allocated to the Ivy Bridge processors and 8 gigabytes to the Xeon Phi processors. Overall, the system holds a total of 1.404 petabytes of memory.<ref name="VisitDongarra"/> |
Tianhe-2 consists of 16,000 compute nodes, each comprising two Intel [[Ivy Bridge (microarchitecture)|Ivy Bridge]] Xeon processors and three [[Intel MIC#Xeon Phi|Xeon Phi]] chips for a total of 3,120,000 cores.<ref name="TOP500June2013"/> This represents the world's biggest installation of Ivy Bridge and Xeon Phi chips. Each of the 16,000 nodes possess 88 gigabytes of memory allocated to the Ivy Bridge processors and 8 gigabytes to the Xeon Phi processors. Overall, the system holds a total of 1.404 petabytes of memory.<ref name="VisitDongarra"/> |
||
During the testing phase, Tianhe-2 was laid out in a non-optimal confined space. When assembled at its final location, the system will have a theoretical peak performance of 54.9 petaflops. At peak power consumption, |
During the testing phase, Tianhe-2 was laid out in a non-optimal confined space. When assembled at its final location, the system will have a theoretical peak performance of 54.9 petaflops. At peak power consumption, the system itself would draw 17.6 megawatts of power, while the external cooling system would draw a further 24 megawatts. The computer complex would occupy 720 square meters of space.<ref name="VisitDongarra"/> |
||
The front-end system consists of 4096 Galaxy FT-1500 CPUs, a [[SPARC]] derivative designed and built by NUDT. Each FT-1500 has 16 cores and a 1.8 gigahertz cycle time. The chip has a performance of 144 gigaflops and runs on 65 watts. The [[interconnect]], called the TH Express-2,designed by NUDT, utilizes a [[fat tree]] [[Network topology|topology]] with 13 switches each of 576 ports.<ref name="VisitDongarra"/> |
The front-end system consists of 4096 Galaxy FT-1500 CPUs, a [[SPARC]] derivative designed and built by NUDT. Each FT-1500 has 16 cores and a 1.8 gigahertz cycle time. The chip has a performance of 144 gigaflops and runs on 65 watts. The [[interconnect]], called the TH Express-2,designed by NUDT, utilizes a [[fat tree]] [[Network topology|topology]] with 13 switches each of 576 ports.<ref name="VisitDongarra"/> |
||
Revision as of 16:01, 23 June 2013
| Sponsors | 863 Program |
|---|---|
| Location | Guangzhou, China |
| Architecture | Intel Xeon, Xeon Phi, Kylin Linux[1] |
| Power | 17.6 MW (24 MW with cooling) |
| Memory | 1,375 TiB (1,000 TiB CPU and 375 TiB Coprocessor)[1] |
| Storage | 12.4 PB |
| Speed | 33.86 PFLOPS |
| Cost | 2.4 billion Yuan (390 million USD)[2] |
| Purpose | Research and education |
Tianhe-2 or TH-2 (Chinese: 天河-2; pinyin: tiānhé-èr; lit. 'Milky Way-2') is a 33.86 petaflop supercomputer located in Guangzhou, China.[3] It is currently the world's fastest supercomputer according to the TOP500 list for June 2013.[4][5]
History
The development of Tianhe-2 was sponsored by the 863 High Technology Program, initiated by the the Chinese Government, the government of Guangdong province, and the government of Guangzhou city.[1] It was built by China's National University of Defense Technology (NUDT) in collaboration with the Chinese IT firm Inspur.[1][5] Inspur manufactured the printed circuit boards and helped with the installation and testing of the system software.[1] The project was originally scheduled for completion in 2015, but was instead declared operational in June 2013.[6] As of June 2013, The Supercomputer has yet to become fully operational. It is expected to reach its full computing capabilities by the end of 2013.[5]
In June 2013, Tianhe-2 topped the TOP500 list of fastest supercomputers in the world. The computer beat out second place finisher Titan by nearly a 2-to-1 margin. Titan, which is housed at the U.S. Department of Energy's Oak Ridge National Laboratory, achieved 17.59 petaflops, while Tianhe-2 achieved 33.86 petaflops. Tianhe-2's phenomenal performance returned the title of the world's fastest supercomputer to China after Tianhe-1's debut in November 2010. The Institute of Electrical and Electronics Engineers said Tianhe-2's win "symbolizes China's unflinching commitment to the supercomputing arms race".[5] China houses 66 of the top 500 supercomputers, second only to the United States' 252 systems.[3]
Also in June 2013, Tianhe-2 is ranked sixth on the Graph500 list of top supercomputers. In their benchmark, the system tested at 2061 giga-TEPS (traversed edges per second). The top system, IBM Sequoia, tested at 15363 giga-TEPS.[4]
Tianhe-2 will be housed at the National Supercomputing Center in Guangzhou (NSCC-GZ) after the completion of the testing procedures.[1]
Applications
According to NUDT, Tianhe-2 will be used for simulation, analysis, and government security applications.[1]
Specifications
Tianhe-2 consists of 16,000 compute nodes, each comprising two Intel Ivy Bridge Xeon processors and three Xeon Phi chips for a total of 3,120,000 cores.[3] This represents the world's biggest installation of Ivy Bridge and Xeon Phi chips. Each of the 16,000 nodes possess 88 gigabytes of memory allocated to the Ivy Bridge processors and 8 gigabytes to the Xeon Phi processors. Overall, the system holds a total of 1.404 petabytes of memory.[1]
During the testing phase, Tianhe-2 was laid out in a non-optimal confined space. When assembled at its final location, the system will have a theoretical peak performance of 54.9 petaflops. At peak power consumption, the system itself would draw 17.6 megawatts of power, while the external cooling system would draw a further 24 megawatts. The computer complex would occupy 720 square meters of space.[1]
The front-end system consists of 4096 Galaxy FT-1500 CPUs, a SPARC derivative designed and built by NUDT. Each FT-1500 has 16 cores and a 1.8 gigahertz cycle time. The chip has a performance of 144 gigaflops and runs on 65 watts. The interconnect, called the TH Express-2,designed by NUDT, utilizes a fat tree topology with 13 switches each of 576 ports.[1]
Tianhe-2 runs on Kylin Linux, a version of the operating system developed by NUDT. Resource management is based on Simple Linux Utility for Resource Management (SLURM).[1]
See also
References
- ^ a b c d e f g h i j k Dongarra, Jack (June 3, 2013). "Visit to the National University for Defense Technology Changsha, China" (PDF). Netlib. Retrieved June 17, 2013.
- ^ Chen, Stephen (20 June 2013). "World's fastest supercomputer may get little use".
- ^ a b c "June 2013". TOP500. Retrieved 17 June 2013.
- ^ a b "The Graph 500 List: June 2013". Graph 500. Retrieved June 19, 2013.
- ^ a b c d Davey Alba (June 17, 2013). "China's Tianhe-2 Caps Top 10 Supercomputers". IEEE Spectrum. Retrieved June 19, 2013.
- ^ Michael Kan, IDG News Service (2012-10-31). "China is building a 100-petaflop supercomputer". infoworld.com. Retrieved 2012-10-31.