Tuesday, April 19, 2011

One ranked first in the world Tianhe - elite





    enhancements to the software, the second system from in terms of peak performance than the previous system, a 3.89-fold increase, more commendable is more continuous performance improvement, increased 4.45 times.
    National Defense University and the Tianjin Binhai New Area through close cooperation, in September last year, the successful development of peak performance per 1,206,000,000,000,mac brushes,000 times as China's first petaflop supercomputer system to realize China has independently developed a super computer from the one hundred trillion times to a petaflop by leaps and bounds.
    then, No.
    It is understood, planning. This marks the completion of the first of a petaflop supercomputer computing centers, scientific research and economic development for our country strong impetus.
    the northern region, the whole of China can be said to have a larger, shared use of the computing platform can be said of our country on the basis of scientific facilities to a new level.

    data for: the current world's fastest computer - the United States Department of Energy's Oak Ridge National Nuclear Laboratory Jaguar (Jaguar),mac makeup, mainly for the simulation and calculation of the current U.S. nuclear test the reliability of nuclear weapons used in storage.

    2010 年 6 months of the world supercomputer list, ranking the first column is the second column for the computer owner, and the third column is the computer name and configuration, the fourth the number of columns for the microprocessor core, the fifth column is the measured maximum computing power, the sixth column is pure performance. Is worth noting that, November last year, No. 5, introduced by the Chinese National Defense University's Milky Way One (563.1TFLOPS), the ranking dropped to No. 7.

    high-performance computers, rack chassis can plug multiple blades, high density),vibram 5 fingers, the blade computing system for each memory, a total of 96G of memory). A group of servers were calculated with 10 Nebula system is based on the dawn of the TC3600's own blade servers, using six-core Intel Xeon X5650 processor, NVIDIA also with a framework based on Fermi's Tesla C2050 GPU acceleration unit, a total of 120,640 cores.

    TC2600 generation blade server based on the dawn of the dawn of 5000 supercomputer


    This is the dawn of the combined company Tianjin base debugging However, hybrid computation, in particular,mac makeup wholesale, the practical application of GPU acceleration are new issues in the world. Therefore, But this represents a future supercomputer development.



    This is the dawn of the combined company Tianjin base debugging