Algorithm-oriented design of efficient many-core architectures applied to dense matrix multiplication

Recent integrated circuit technologies have opened the possibility to design parallel architectures with hundreds of cores on a single chip. The design space of these parallel architectures is huge with many architectural options. Exploring the design space gets even more difficult if, beyond performance and area, we also consider extra metrics like performance and area efficiency, where the designer tries to design the architecture with the best performance per chip area and the best sustainable performance. In this paper we present an algorithm-oriented approach to design a many-core architecture. Instead of doing the design space exploration of the many core architecture based on the experimental execution results of a particular benchmark of algorithms, our approach is to make a formal analysis of the algorithms considering the main architectural aspects and to determine how each particular architectural aspect is related to the performance of the architecture when running an algorithm or set of algorithms. The architectural aspects considered include the number of cores, the local memory available in each core, the communication bandwidth between the many-core architecture and the external memory and the memory hierarchy. To exemplify the approach we did a theoretical analysis of a dense matrix multiplication algorithm and determined an equation that relates the number of execution cycles with the architectural parameters. Based on this equation a many-core architecture has been designed. The results obtained indicate that a 100 mm(2) integrated circuit design of the proposed architecture, using a 65 nm technology, is able to achieve 464 GFLOPs (double precision floating-point) for a memory bandwidth of 16 GB/s. This corresponds to a performance efficiency of 71 %. Considering a 45 nm technology, a 100 mm(2) chip attains 833 GFLOPs which corresponds to 84 % of peak performance These figures are better than those obtained by previous many-core architectures, except for the area efficiency which is limited by the lower memory bandwidth considered. The results achieved are also better than those of previous state-of-the-art many-cores architectures designed specifically to achieve high performance for matrix multiplication.

A Compact and Scalable RNS Architecture

Conferência: IEEE 24th International Conference on Application-Specific Systems, Architectures and Processors (ASAP)- Jun 05-07, 2013

Electrodeposition and characterization of nickel–copper metallic foams for application as electrodes for supercapacitors

Nickel-copper metallic foams were electrodeposited from an acidic electrolyte, using hydrogen bubble evolution as a dynamic template. Their morphology and chemical composition was studied by scanning electron microscopy and related to the deposition parameters (applied current density and deposition time). For high currents densities (above 1 A cm(-2)) the nickel-copper deposits have a three-dimensional foam-like morphology with randomly distributed nearly-circular pores whose walls present an open dendritic structure. The nickel-copper foams are crystalline and composed of pure nickel and a copper-rich phase containing nickel in solid solution. The electrochemical behaviour of the material was studied by cyclic voltammetry and chronopotentiometry (charge-discharge curves) aiming at its application as a positive electrode for supercapacitors. Cyclic voltammograms showed that the Ni-Cu foams have a pseudocapacitive behaviour. The specific capacitance was calculated from charge-discharge data and the best value (105 F g(-1) at 1 mA cm(-2)) was obtained for nickel-copper foams deposited at 1.8 A cm(-2) for 180 s. Cycling stability of these foams was also assessed and they present a 90 % capacitance retention after 10,000 cycles at 10 mA cm(-2).

Electrodeposition and characterization of nikel-copper metallic foams for application as electrodes for supercapacitors

Nickel-copper metallic foams were electrodeposited from an acidic electrolyte, using hydrogen bubble evolution as a dynamic template. Their morphology and chemical composition was studied by scanning electron microscopy and related to the deposition parameters (applied current density and deposition time). For high currents densities (above 1 A cm(-2)) the nickel-copper deposits have a three-dimensional foam-like morphology with randomly distributed nearly-circular pores whose walls present an open dendritic structure. The nickel-copper foams are crystalline and composed of pure nickel and a copper-rich phase containing nickel in solid solution. The electrochemical behaviour of the material was studied by cyclic voltammetry and chronopotentiometry (charge-discharge curves) aiming at its application as a positive electrode for supercapacitors. Cyclic voltammograms showed that the Ni-Cu foams have a pseudocapacitive behaviour. The specific capacitance was calculated from charge-discharge data and the best value (105 F g(-1) at 1 mA cm(-2)) was obtained for nickel-copper foams deposited at 1.8 A cm(-2) for 180 s. Cycling stability of these foams was also assessed and they present a 90 % capacitance retention after 10,000 cycles at 10 mA cm(-2).

Explicit series solution for a glucose-induced electrical activity model of pancreatic beta-cells

In this work, we present the explicit series solution of a specific mathematical model from the literature, the Deng bursting model, that mimics the glucose-induced electrical activity of pancreatic beta-cells (Deng, 1993). To serve to this purpose, we use a technique developed to find analytic approximate solutions for strongly nonlinear problems. This analytical algorithm involves an auxiliary parameter which provides us with an efficient way to ensure the rapid and accurate convergence to the exact solution of the bursting model. By using the homotopy solution, we investigate the dynamical effect of a biologically meaningful bifurcation parameter rho, which increases with the glucose concentration. Our analytical results are found to be in excellent agreement with the numerical ones. This work provides an illustration of how our understanding of biophysically motivated models can be directly enhanced by the application of a newly analytic method.