Description

Some of the baskets can be replaced with special nodes with graphic accelerators (AMD Firestream, NVIDIA Tesla, in the future NVIDIA Fermi);

Peak performance of universal computing unit processors is up to 249.6 GFlops (up to 364.8 Gflops in the first quarter). One processor - a 6-core AMD Istanbul microprocessor (hereinafter - with 12-core AMD Magny-Cors and 8-core multi-thread Intel Nehalem-EX); single-node RAM up to 256 GB, globally addressable within the node;

Peak basket performance - 2.5-3.6 Tflops, power consumption ~ 6 kW;

Peak rack performance - 12-18 Tflops, power consumption ~ 30 kW;

Peak node performance with GPGPU at double precision up to 1.6 TFlops (up to 3 TFlops in the second quarter);

Built-in in a rack or external disk array with a capacity of 14 TB to dozens of PB with support for Global Name Space;

InfiniBand DDR, QDR, 10 Gbps Ethernet, 1 Gbps Ethernet interconnect networks; Engineering infrastructure of power supply and liquid or air cooling, RAS subsystem (availability, serviceability, reliability);

Standard software (Linux, Windows, C, C ++, Fortran, MPI, OpenMP), RDMA libraries (GASNet, ARMCI), parallel compilers DVM-C and DVM-FORTRAN (original Russian developments), parallel programming language Charm ++ based on computational models with message-driven control of messages, parallel programming languages of the PGAS class (UPC, Co-Array Fortran), as well as the latest promising languages Chapel and X10; Installation and maintenance of application packages of Russian and foreign production programs for solving problems in mechanical engineering, biology and pharmaceutics - Marc, Nastran, Ansys, LS-Dyna, Star-CD, Fluent, FlowVision (Russia), ASONIKA (Russia). In addition, it is possible to install a special library, an original development, for efficient work with unstructured data and development of information systems, as well as a library for solving complex computational problems and data analysis tasks;

Remote administration system with 24/7 support; Development of parallel applications for customers.