Приближенный алгоритм выбора оптимального подмножества узлов в коммуникационной сети Ангара с отказами
The Angara high-speed interconnect with multidimensional torus topology is under development in Scientific Research Center for Electronic Computer Technology. During the utilization of the Angara interconnect in cluster systems, there exist busy and failed nodes. Thus, there is a problem of finding an optimal cluster node subset such that the network traffic belongs to this node subset and the node subset size is not less than a given size. The paper presents an approximate algorithm for solving this problem.
Increasing evidence suggests that neuronal communication is a defining property of functionally specialized brain networks and that it is implemented through synchronization between population activities of distinct brain areas. The detection of long-range coupling in electroencephalography (EEG) and magnetoencephalography (MEG) data using conventional metrics (such as coherence or phase-locking value) is by definition contaminated by spatial leakage. Methods such as imaginary coherence, phase-lag index or orthogonalized amplitude correlations tackle spatial leakage by ignoring zero-phase interactions. Although useful, these metrics will by construction lead to false negatives in cases where true zero-phase coupling exists in the data and will underestimate interactions with phase lags in the vicinity of zero. Yet, empirically observed neuronal synchrony in invasive recordings indicates that it is not uncommon to find zero or close-to-zero phase lag between the activity profiles of coupled neuronal assemblies. Here, we introduce a novel method that allows us to mitigate the undesired spatial leakage effects and detect zero and near zero phase interactions. To this end, we propose a projection operation that operates on sensor-space cross-spectrum and suppresses the spatial leakage contribution but retains the true zero-phase interaction component. We then solve the network estimation task as a source estimation problem defined in the product space of interacting source topographies. We show how this framework provides reliable interaction detection for all phase-lag values and we thus refer to the method as Phase Shift Invariant Imaging of Coherent Sources (PSIICOS). Realistic simulations demonstrate that PSIICOS has better detector characteristics than existing interaction metrics. Finally, we illustrate the performance of PSIICOS by applying it to real MEG dataset recorded during a standard mental rotation task. Taken together, using analytical derivations, data simulations and real brain data, this study presents a novel source-space MEG/EEG connectivity method that overcomes previous limitations and for the first time allows for the estimation of true zero-phase coupling via non-invasive electrophysiological recordings.
Accuracy of aircraft district location determination in changing daily, weather climatic conditions is a crucial factor which civil and military aviation usage efficiency and sometimes appropriateness is dependent on. Despite permanent inertial and satellite navigating systems, which all modern aircrafts are equipped, by performance quality improvement still required accuracy of aircraft district location determination is not reached. This fact still forces to use three-dimensional navigating systems which are installed aboard of specialized aircrafts for civil applications and which are essential for military aircraft of 5th generation. Latest three-dimensional navigating systems investigations show that their performance accuracy improvement requires joint use of methods and algorithms of estimation, identification and test of hypothesis during stereoscopic pairs of large ground images real-time processing. Therefore success of three-dimensional navigating systems practical usage is mainly defined by performance, computing stability and fault tolerance of onboard computing systems which are dominantly based on RISC-processors and FPGA-video-processors (Field Programmable Gate Array). The received empirical estimations of required performance, computing stability and fault tolerance show: 1. The elementary searchless district location determination based on one stereopair processing requires ~(0,12-0,6) *108 operations of multiplication and addition. Such a task is able to be processed in 1 second by RISC-processor with performance not below some Gflops, taking into account computing organization system time costs. 2. If micro program management intermediate level is presented in FPGA-technologies based on VHDL than algorithmically-focused MIMD-bit-stream three-dimensional navigating systems videoprocessor is potentially implementable on the basis of one VLSI FPGA Xilinx Virtex 6. Such a processor is possible to achieve performance of V=FT*Lp/n=500*106*192/64≈1,5*109 operations per second in double-format. This performance rate is enough for target size stereopairs processing rate of 25÷125 images per second so given performance exceeds algorithmic costs for one required format stereopair processing exactly this amount of times. Here: FT - bit-matrix clock rate, Lp - one cycle operations quantity, n - arithmetics capacity. It is fundamental that there are no system time costs on real-time computing resources management in MIMD-bit-stream computing technologies so that computing expenses become comparable with algorithmic in such videoprocessors. 3. Despite the fact that required by three-dimensional navigating systems algorithms computing performance and accuracy are achievable both in RISC-architecture and specialized FPGA-based architectures, 5th generation aircrafts such as F-35 should stabilize flight parameters, used for high-precision own location determination, which is realizable only at its separate stages. At the same time it should be kept in mind that computing paralleling index increase more than 4 times may lead to loss of onboard three-dimensional navigating system computer based on RISC-platforms computing stability. 4. RISC-architecture fault tolerance more than 10 times less than specialized FPGA-bit-stream architectures one. It is the most essential for spaceships which lifetime should exceed VLSI components nonfailure operating time more than 10 times. Such fault tolerance indicators qualitative divergence is caused by: - indifferent failures presence in bit-stream architectures which quantity exceeds ratio of failures parried by RISC-architecture processors majority reservation schemes; - possibility of ~112 times bit-stream microprograms field re-allocation on a bit matrix tolerantly operating failure map.
The analysis of mesh, torus and irregular topologies, revealing their advantages and disadvantages to build networks‑on‑chip (NoC) is given. It is proposed the use of two-dimensional circulant networks and the routing method, which makes it possible to improve the characteristics of NoC topology up to 40%.