High-performance computing (HPC) is recognized as one of the pillars for further progress in science, industry, medicine, and education. Current HPC systems are being developed to overcome emerging architectural challenges in order to reach Exascale level of performance, projected for the year 2020. The much larger embedded and mobile market allows for rapid development of intellectual property (IP) blocks and provides more flexibility in designing an application-specific system-on-chip (SoC), in turn providing the possibility in balancing performance, energy-efficiency, and cost. In the Mont-Blanc project, we advocate for HPC systems being built from such commodity IP blocks, currently used in embedded and mobile SoCs. As a first demonstrator of such an approach, we present the Mont-Blanc prototype; the first HPC system built with commodity SoCs, memories, and network interface cards (NICs) from the embedded and mobile domain, and off-the-shelf HPC networking, storage, cooling, and integration solutions. We present the system's architecture and evaluate both performance and energy efficiency. Further, we compare the system's abilities against a production level supercomputer. At the end, we discuss parallel scalability and estimate the maximum scalability point of this approach across a set of applications.

This paper provides an asymptotic (in the number of snapshots) closed form expression of the bias and RMS (root mean square) error of the estimated DOA (direction of arrival) for the algorithm recently introduced in Ferreol et al. (2004). This algorithm provides a 1D DOA estimation in a multi-parameter context where the DOA have to be estimated with some separable nuisance parameters. Results are based on a second order approximation of the criterion. DOA estimation errors are then expressed as a ratio of Hermitian forms of multivariate complex random variables. Theoretical results are validated by simulations in a self-calibration context.

In array processing, the far-field assumption of planar wavefronts is widely used by direction of arrival (DOA) estimators but not always satisfied. In this letter, we introduce a new method for bearing-range estimation, which extends classical subspace-based bearing estimators to a curved wavefront context. The bearing estimation is provided by a one-dimensional procedure, and ranges are simply analytically deduced from the bearing estimates. The proposed approach is illustrated by the introduction of the music curved wavefront (MCW) algorithm.

In estimation theory, the asymptotic efficiency of the maximum likelihood (ML) method for independent identically distributed observations and when the number of observations, T, tends to infinity is a well known result. In some scenarios, the number of snapshots may be small, making this result inapplicable. In the array processing framework, for Gaussian emitted signals, we fill this lack at high signal-to-noise ratio (SNR). In this situation, we show that the ML estimation is asymptotically (with respect to SNR) inefficient and non Gaussian.

The purpose of this letter is to investigate, in the deterministic case, the nonasymptotic behavior of beamforming estimation of the direction of arrival of a single source impinging on a uniform linear array. We derive an approximate analytical expression of the maximum-likelihood mean-squared error that is valid for all SNR ranges and number of snapshots. Computer simulations confirm the validity of the theoretical investigations.

This letter is devoted to the analysis of the nonasymptotic behavior of beamforming for the bearing estimation of a single stochastic source impinging on an uniform linear array. An analytic expression of the maximum-likelihood mean-squared error is derived. This general expression is valid at any SNR and for any number of snapshots. Simulation results verify the analytically predicted performances.

It is well known that the stochastic maximum-likelihood (SML) method yields Gaussian and efficient estimates when the number of independent identically distributed samples tends to infinity. This letter investigates the behavior of SML bearing estimation for a single source impinging on an antenna array when the SNR tends to infinity for a fixed number of samples. We prove that, rather surprisingly, the bearing estimates are not efficient and are asymptotically distributed according to a Student law. Simulation results confirm theoretical analysis.

The purpose of this work is the estimation of Doppler echoes spectral moments. In case of strong overlapping, Fourier-like techniques provide poor results because of the lack of resolution. We propose the use of stochastic maximum-likelihood (SML) and subspace-based methods (WPSF algorithm) for a joint estimation of spectral moments. The statistical performances (theoretical and empirical by Monte Carlo simulations) of estimators are compared with the Cramer-Rao lower bound. The results of tests performed on very high frequency (VHF) times series obtained during Thunderstorm, Arecibo, PR during September and October 1998 validate the model and algorithms and confirm the interest of both approaches.