June 17, 2013
LEIPZIG, Germany, June 17 -- Marking a new era in supercomputing, the MilkyWay-2 system today was recognized as the most powerful supercomputer in the world according to top500.org. The system registered a performance of 33.86PF on the industry standard Linpack benchmark on a system with 16000 compute nodes, including over 32000 of upcoming Intel Xeon processor E5-2600 v2 and over 48000 Intel Xeon Phi coprocessors. The total system power consumption was 17.8 MW, which made the system one of the most power efficient systems on the top 500 list with a performance per watt of over 1.9 GF/W.
The system was designed and developed by the National University of Defense Technology (NUDT) and uses what is called a “neo-heterogeneous architecture.” The previous TianHe-1A system was built using a fully heterogeneous architecture combining traditional CPUs with attached accelerators. While this architecture made great strides in energy efficiency, in this architecture, the CPUs and accelerators use a very different programming model, which makes the development of scientific applications more complex and time consuming. In a “neo-heterogeneous architecture” while the hardware architecture has multiple classes of hardware compute capabilities, they are all accessed by a common programming model, streamlining development and optimization processes.
The MilkyWay-2 system is based on a NUDT designed proprietary high-speed network called TH Express-2, and has over 1 Petabytes of memory. It has over 12 Petabytes of global shared parallel storage, and includes 125 cabinets for computation, 13 cabinets for communication and 24 cabinets for storage.
The MilkyWay-2 system is the largest All-Intel based system ever built. It is among the first to use the future Intel Xeon Processor E5-2600 V2 processors, and the largest to-date installation of Intel Xeon Phi coprocessors.
In MilkyWay-2, the proprietary high-speed interconnection network is provided, which adopts the opto-electronic hybrid interconnection and the fat-tree topology. With the bi-directional bandwidth at 160Gbps, the network itself is balanced and scalable.
The service array of MilkyWay-2 was built with the FT-1500 processors. The FT-1500 processor, integrating 16 quad-threaded cores, implements the vector computational extensions for large-scale scientific and engineering applications. The chip runs at a speed of up to 1.8GHz with a peak performance of 144GFlops, while consuming 65W of power on average.
MilkyWay-2 designs and implements the new hierarchical accelerated storage architecture. The architecture combines the distributed and the shared storage and implements a parallel hybrid hierarchical file system H2FS, fits to the big data management and the storage optimization, which also supports various storage semantics.
A new parallel programming model OpenMC is introduced in MilkyWay-2, which improves the programmability of applications. Via a unified abstraction for CPU and accelerator, the OpenMC model takes full advantage of the neo-heterogeneous structure. Meanwhile, the OpenMC efficiently exploit multi-level parallelisms through optimizations such as network topology based job deployment, adaptive load balance and communication hiding.
In MilkyWay-2, the compute modules are inserted symmetrically from both front and rear sides into the middle backplane while one CPM blade and one APU blade are horizontally blind connected, ensuring the higher computing density. There are 64 compute modules including 256 CPUs and 384 accelerators installed in one single cabinet. The computing density of single cabinet reaches 439.3TFlops.
As a result of above innovative designs, this benchmark was delivered in record time. The first Intel Xeon Processor E5-2600 V2 based nodes were delivered by Inspur in early April 2013. The system was integrated, debugged and the benchmark was run on the full system in about 8 weeks. A remarkable technical achievement in its own right brought upon by NUDT professional mainboard designing, system construction technology and the strengths of a neo-heterogeneous architecture.
This system was sponsored by 863 High tech. Program of the Ministry of Science and Technology of China, the Government of Guangdong province and the Government of Guangzhou city, operated by NUDT, this system is intended to serve as an open platform for research and education and also intended to provide high performance computing service for southern China.
MilkyWay-2 system will be applied to research a wide range of scientific applications, including finding the ability to more accurately predict weather patterns, create more efficient energy resources, and develop cures for diseases, among numerous other applications. In addition, it expects to work in the field of “Big Data applications” to generate the ability to quickly compute, simulate and make more informed decisions.
National University of Defense Technology (NUDT) is located in Changsha, the capital city of Hunan Province in south-central China. The university is mainly for science and engineering education. It is designated to Project 211 and Project 985, the two national plans for facilitating the development of Chinese higher education. Now the university consists of 10 colleges, under which there are over 40 departments, institutes and national laboratories.
As the first department of computer science in China, the Department of Computer Science in NUDT was established in 1966. It was expanded and renamed as the College of Computer Science in 1999. In the National General Evaluation of computer subject in 2002 and 2006, the subject of computer science and technology was ranked NO.1 and NO.2 in China. In 2012, the subject of software engineering was ranked No.1 and the subject of computer science and technology was ranked No.2.
Since 1978, NUDT has begun to research and build supercomputers. Throughout the past three decades, NUDT has achieved brilliant innovations in the field of high performance computing. A series of supercomputers were developed by NUDT, and most of them reached the highest performance in China and the world.
10/30/2013 | Cray, DDN, Mellanox, NetApp, ScaleMP, Supermicro, Xyratex | Creating data is easy… the challenge is getting it to the right place to make use of it. This paper discusses fresh solutions that can directly increase I/O efficiency, and the applications of these solutions to current, and new technology infrastructures.
10/01/2013 | IBM | A new trend is developing in the HPC space that is also affecting enterprise computing productivity with the arrival of “ultra-dense” hyper-scale servers.
Ken Claffey, SVP and General Manager at Xyratex, presents ClusterStor at the Vendor Showdown at ISC13 in Leipzig, Germany.
Join HPCwire Editor Nicole Hemsoth and Dr. David Bader from Georgia Tech as they take center stage on opening night at Atlanta's first Big Data Kick Off Week, filmed in front of a live audience. Nicole and David look at the evolution of HPC, today's big data challenges, discuss real world solutions, and reveal their predictions. Exactly what does the future holds for HPC?