Recently, the manufactures of supercomputers have made use of FPGAs to accelerate scientific applications [16] [17]. Traditionally, the FPGAs were used only on non-scientific applications. The main reasons for this fact are: the floating-point computation complexity; the FPGA logic cells are not sufficient for the scientific cores implementation; the cores complexity prevents them to operate on high frequencies.Nowadays, the increase of specialized blocks availability in complex operations, as sum and multiplier blocks, implemented directly in FPGA and, the increase of internal RAM blocks (BRAMs) have made possible high performance systems that use FPGA as a processing element for scientific computation [2].These devices are used as co-processors that execute intensive computation. The emphasis of these architectures is the exploration of parallelism present on scientific computation operations and data reuse.In major of these applications, the scientific computation uses, in general, operations of big floating-point dense matrices, which are normally operated by MACs.In this work, we describe the architecture of an accumulative multiplier (MAC) in double precision floating-point, according to IEEE-754 standard and we propose the architecture of a multiplier of matrices that uses developed instances of the MACs and explores the reuse of data through the use of the BRAMs (Blocks of RAM internal to the FPGAs) of a Xilinx Virtex 4 LX200 FPGA. The synthesis results showed that the implemented MAC could reach a performance of 4GFLOPs.

No abstract

No abstract

A computational Bayesian framework for inference regarding the range, depth and velocity of a submerged mobile localized body in an ocean waveguide is constructed. The approach incorporates the various mixed mode Doppler-frequency dispersion effects associated with the reverberant body’s vertical angle scattering. Such coupled eigen returns present angle-frequency modes of intermediate Doppler residing between that of the two coupled specular eigen path Doppler frequencies. It is exactly these modes which present a severe limitation to exploiting closely spaced arrivals in coherence and aperture limited environments. Conditional densities of arrival angle-Doppler are solved via a fast inverse quantile sampler. All other conditionals offer closed form representations in the Gaussian-inverse gamma family. The joint posterior probability density (PPD) of the arrivals are numerically solved and the implied PPD of the object's range, depth, and speed is inferred through acoustic ray interpolation. Case studies are presented with various refractive ocean waveguide environments as well as iso-velocity cases. The framework offers a means of incorporating spatio-temporal arrival structure for recursive tracking in an active sonar system. [Work is funded by the Office of Naval Research.]

Considered here is an narrow band directed source and hydrophone receiver arrangement employed to infer the depth, speed, and range of an oncoming submerged object. Tracking the scattering body by means of a continuous wave transmission is challenging due to the difficulty of inferring the frequencies and angles of the two returned closely spaced wave vectors. Computation of the posterior pdf of these two wave vectors is accomplished by a judicious Gibbs sampling scheme that accounts for the uncertainty in the ambient acoustic noise level. Computational improvements are accomplished by taking full advantage of the prior distribution of the wave vectors associated with the specific target scenario. Very short duration observations of approximately 10 milliseconds are considered over which the Doppler rate of change of the two wave vectors can be considered negligible. This Bayesian scheme takes advantage of the analytic tractability of the conditional density of the received amplitudes and phases and of the noise powers. The conditional densities of the ordered wave vectors however are constructed numerically by 2 dimensional inverse quantile sampling. The inferred joint density of depth, range, and speed of the target is accomplished by constructing an inverse transformation of the acoustic propagation model.