A System-C based Micro architectural Exploration Framework for Latency, Power and Performance Trade-offs of On-Chip Interconnection

We describe a System-C based framework we are developing, to explore the impact of various architectural and microarchitectural level parameters of the on-chip interconnection network elements on its power and performance. The framework enables one to choose from a variety of architectural options like topology, routing policy, etc., as well as allows experimentation with various microarchitectural options for the individual links like length, wire width, pitch, pipelining, supply voltage and frequency. The framework also supports a flexible traffic generation and communication model. We provide preliminary results of using this framework to study the power, latency and throughput of a 4x4 multi-core processing array using mesh, torus and folded torus, for two different communication patterns of dense and sparse linear algebra. The traffic consists of both Request-Response messages (mimicing cache accesses)and One-Way messages. We find that the average latency can be reduced by increasing the pipeline depth, as it enables higher link frequencies. We also find that there exists an optimum degree of pipelining which minimizes energy-delay product.

Online Unsupervised Pattern Discovery in Speech Using Parallelization

Segmental dynamic time warping (DTW) has been demonstrated to be a useful technique for finding acoustic similarity scores between segments of two speech utterances. Due to its high computational requirements, it had to be computed in an offline manner, limiting the applications of the technique. In this paper, we present results of parallelization of this task by distributing the workload in either a static or dynamic way on an 8-processor cluster and discuss the trade-offs among different distribution schemes. We show that online unsupervised pattern discovery using segmental DTW is plausible with as low as 8 processors. This brings the task within reach of today's general purpose multi-core servers. We also show results on a 32-processor system, and discuss factors affecting scalability of our methods.

Structure and Bonding in Stannadiphospholes and their Dianions SnC(2)P(2)R(2)(m) (R=H, tBu m=0,-2): A Comparative Study with C(5)H(5)(+) and C(5)H(5)(-) Analogues

The potential energy surfaces of both neutral and dianionic SnC(2)P(2)R(2) (R=H, tBu) ring systems have been explored at the B3PW91/LANL2DZ (Sn) and 6-311 + G* (other atoms) level. In the neutral isomers the global minimum is a nido structure in which a 1,2-diphosphocyclobutadiene ring (1,2-DPCB) is capped by the Sn. Interestingly, the structure established by Xray diffraction analysis, for R=tBu, is a 1,3-DPCB ring capped by Sn and it is 2.4 kcal mol(-1) higher in energy than the 1,2-DPCB ring isomer. This is possibly related to the kinetic stability of the 1,3-DPCB ring, which might originate from the synthetic precursor ZrCp(2)tBu(2)C(2)P(2). In the case of the dianionic isomers we observe only a 6 pi-electron aromatic structure as the global minimum, similarly to the cases of our previously reported results with other types of heterodiphospholes.([1,4,19]) The existence of large numbers of cluster-type isomers in neutral and 6 pi-planar structures in the dianions SnC(2)P(2)R(2)(2-) (R=H, tBu) is due to 3D aromaticity in neutral clusters and to 2D pi aromaticity of the dianionic rings. Relative energies of positional isomers mainly depend on: 1) the valency and coordination number of the Sn centre, 2) individual bond strengths, and 3) the steric effect of tBu groups. A comparison of neutral stannadiphospholes with other structurally related C(5)H(5)(+) analogues indicates that Sn might be a better isolobal analogue to P(+) than to BH or CH(+). The variation in global minima in these C(5)H(5)(+) analogues is due to characteristic features such as 1) the different valencies of C, B, P and Sn, 2) the electron deficiency of B, 3) weaker p pi-p pi bonding by P and Sn atoms, and 4) the tendency of electropositive elements to donate electrons to nido clusters. Unlike the C5H5+ systems, all C(5)H(5)(-) analogues have 6 pi-planar aromatic structures as global minima. The differences in the relative ordering of the positional isomers and ligating properties are significant and depend on 1) the nature of the pi orbitals involved, and 2) effective overlap of orbitals.

Infrared studies of the phase transitions of alkylammonium halides, RNH3X, and bis-(alkylammonium) tetrahalogenometallates(II), (RNH3)2MX4, (R = alkyl, M = metal, X = Cl or Br)

Internal vibration modes of bis-(alkylammonium) tetrachlorometallates(II) and the corresponding alkylammonium chlorides have been studied through their phase transitions using infrared spectroscopy. The studies show that the vibrational states of alkylammonium ions change markedly through the phase transitions. Spectra of the analogous tetrabromometallates and alkylammonium bromides also confirm this behaviour. There is appreciable motion of the alkylammonium ions in the high-temperature phases; thus, CH3NH+3 ions are essentially undistorted in these phases. The low-temperature, ordered phases show evidence of stronger hydrogen bonding of the cations and for the presence of C—N torsional modes.

Scalable Context-Sensitive Points-To Analysis Using Multi-Dimensional Bloom Filters

Context-sensitive points-to analysis is critical for several program optimizations. However, as the number of contexts grows exponentially, storage requirements for the analysis increase tremendously for large programs, making the analysis non-scalable. We propose a scalable flow-insensitive context-sensitive inclusion-based points-to analysis that uses a specially designed multi-dimensional bloom filter to store the points-to information. Two key observations motivate our proposal: (i) points-to information (between pointer-object and between pointer-pointer) is sparse, and (ii) moving from an exact to an approximate representation of points-to information only leads to reduced precision without affecting correctness of the (may-points-to) analysis. By using an approximate representation a multi-dimensional bloom filter can significantly reduce the memory requirements with a probabilistic bound on loss in precision. Experimental evaluation on SPEC 2000 benchmarks and two large open source programs reveals that with an average storage requirement of 4MB, our approach achieves almost the same precision (98.6%) as the exact implementation. By increasing the average memory to 27MB, it achieves precision upto 99.7% for these benchmarks. Using Mod/Ref analysis as the client, we find that the client analysis is not affected that often even when there is some loss of precision in the points-to representation. We find that the NoModRef percentage is within 2% of the exact analysis while requiring 4MB (maximum 15MB) memory and less than 4 minutes on average for the points-to analysis. Another major advantage of our technique is that it allows to trade off precision for memory usage of the analysis.

MODLEX: A Multi Objective Data Layout EXploration Framework for Embedded Systems-on-Chip

The memory subsystem is a major contributor to the performance, power, and area of complex SoCs used in feature rich multimedia products. Hence, memory architecture of the embedded DSP is complex and usually custom designed with multiple banks of single-ported or dual ported on-chip scratch pad memory and multiple banks of off-chip memory. Building software for such large complex memories with many of the software components as individually optimized software IPs is a big challenge. In order to obtain good performance and a reduction in memory stalls, the data buffers of the application need to be placed carefully in different types of memory. In this paper we present a unified framework (MODLEX) that combines different data layout optimizations to address the complex DSP memory architectures. Our method models the data layout problem as multi-objective genetic algorithm (GA) with performance and power being the objectives and presents a set of solution points which is attractive from a platform design viewpoint. While most of the work in the literature assumes that performance and power are non-conflicting objectives, our work demonstrates that there is significant trade-off (up to 70%) that is possible between power and performance.

Fermion Electric Dipole Moments in R-parity violating Supersymmetry

In this talk I discuss some aspects of the study of electric dipole moments (EDMs) of the fermions, in the context of R-parity violating (\rpv) Supersymmetry (SUSY). I will start with a brief general discussion of how dipole moments, in general, serve as a probe of physics beyond the Standard Model (SM) and an even briefer summary of \rpv SUSY. I will follow by discussing a general method of analysis for obtaining the leading fermion mass dependence of the dipole moments and present its application to \rpv SUSY case. Then I will summarise the constraints that the analysis of $e,n$ and $Hg$ EDMs provide for the case of trilinear \rpv SUSY couplings and make a few comments on the case of bilinear \rpv, where the general method of analysis proposed by us does not work.

Multilevel inverters for low-power application

Multilevel inverters are an attractive solution in the medium-voltage and high-power applications. However in the low-power range also it can be a better solution compared to two-level inverters, if MOSFETs are used as devices switching in the order of 100 kHz. The effect of clamping diodes in the diode-clamped multilevel inverters play an important role in determining its efficiency. Power loss introduced by the reverse recovery of MOSFET body diode prohibits the use of MOSFET in hard-switched inverter legs. A technique of avoiding reverse recovery loss of MOSFET body diode in a three-level neutral point clamped inverter is suggested. The use of multilevel inverters topology enables operation at high switching frequency without sacrificing efficiency. High switching frequency of operation reduces the output filter requirement, which in turn helps in reducing the size of the inverter. This study elaborates the trade-off analysis to quantify the suitability of multilevel inverters in the low-power applications. Advantages of using a MOSFET-based three-level diode-clamped inverter for a PM motor drive and UPS systems are discussed.

The R-x, R2-x Planes; Some New Ideas for the Study of Nonlinear Systems

Use of some new planes such as the R-x, R2-x (where R represents in the n-dimensional phase space, the radius vector from the origin to any point on the trajectory described by the system) is suggested for analysis of nonlinear systems of any kind. The stability conditions in these planes are given. For easy understanding of the method, the transformation from the phase plane to the R-x, R2-x planes is brought out for second-order systems. In general, while these planes serve as useful as the phase plane, they have proved to be simpler in determining quickly the general behavior of certain classes of second-order nonlinear systems. A chart and a simple formula are suggested to evaluate time easily from the R-x and R2-x trajectories, respectively. A means of solving higher-order nonlinear systems is also illustrated. Finally, a comparative study of the trajectories near singular points on the phase plane and on the new planes is made.

Trade determination in multi-attribute exchanges

Electronic exchanges are double-sided marketplaces that allow multiple buyers to trade with multiple sellers, with aggregation of demand and supply across the bids to maximize the revenue in the market. Two important issues in the design of exchanges are (1) trade determination (determining the number of goods traded between any buyer-seller pair) and (2) pricing. In this paper we address the trade determination issue for one-shot, multi-attribute exchanges that trade multiple units of the same good. The bids are configurable with separable additive price functions over the attributes and each function is continuous and piecewise linear. We model trade determination as mixed integer programming problems for different possible bid structures and show that even in two-attribute exchanges, trade determination is NP-hard for certain bid structures. We also make some observations on the pricing issues that are closely related to the mixed integer formulations.