Accelerating Numerical Linear Algebra Kernels on a Scalable Run Time Reconfigurable Platform

Numerical Linear Algebra (NLA) kernels are at the heart of all computational problems. These kernels require hardware acceleration for increased throughput. NLA Solvers for dense and sparse matrices differ in the way the matrices are stored and operated upon although they exhibit similar computational properties. While ASIC solutions for NLA Solvers can deliver high performance, they are not scalable, and hence are not commercially viable. In this paper, we show how NLA kernels can be accelerated on REDEFINE, a scalable runtime reconfigurable hardware platform. Compared to a software implementation, Direct Solver (Modified Faddeev's algorithm) on REDEFINE shows a 29X improvement on an average and Iterative Solver (Conjugate Gradient algorithm) shows a 15-20% improvement. We further show that solution on REDEFINE is scalable over larger problem sizes without any notable degradation in performance.

Editorial optical computing circuits, devices and systems

Information forms the basis of modern technology. To meet the ever-increasing demand for information, means have to be devised for a more efficient and better-equipped technology to intelligibly process data. Advances in photonics have made their impact on each of the four key applications in information processing, i.e., acquisition, transmission, storage and processing of information. The inherent advantages of ultrahigh bandwidth, high speed and low-loss transmission has already established fiber-optics as the backbone of communication technology. However, the optics to electronics inter-conversion at the transmitter and receiver ends severely limits both the speed and bit rate of lightwave communication systems. As the trend towards still faster and higher capacity systems continues, it has become increasingly necessary to perform more and more signal-processing operations in the optical domain itself, i.e., with all-optical components and devices that possess a high bandwidth and can perform parallel processing functions to eliminate the electronic bottleneck.

Passive vibration isolation of reaction wheel disturbances using a low frequency flexible space platform

Reaction wheel assemblies (RWAs) are momentum exchange devices used in fine pointing control of spacecrafts. Even though the spinning rotor of the reaction wheel is precisely balanced to minimize emitted vibration due to static and dynamic imbalances, precision instrument payloads placed in the neighborhood can always be severely impacted by residual vibration forces emitted by reaction wheel assemblies. The reduction of the vibration level at sensitive payloads can be achieved by placing the RWA on appropriate mountings. A low frequency flexible space platform consisting of folded continuous beams has been designed to serve as a mount for isolating a disturbance source in precision payloads equipped spacecrafts. Analytical and experimental investigations have been carried out to test the usefulness of the low frequency flexible platform as a vibration isolator for RWAs. Measurements and tests have been conducted at varying wheel speeds, to quantify and characterize the amount of isolation obtained from the reaction wheel generated vibration. These tests are further extended to other variants of similar design in order to bring out the best isolation for given disturbance loads. Both time and frequency domain analysis of test data show that the flexible beam platform as a mount for reaction wheels is quite effective and can be used in spacecrafts for passive vibration control. (C) 2011 Elsevier Ltd. All rights reserved.

Unzipping and binding of small interfering RNA with single walled carbon nanotube: A platform for small interfering RNA delivery

In an effort to design efficient platform for siRNA delivery, we combine all atom classical and quantum simulations to study the binding of small interfering RNA (siRNA) by pristine single wall carbon nanotube (SWCNT). Our results show that siRNA strongly binds to SWCNT surface via unzipping its base-pairs and the propensity of unzipping increases with the increase in the diameter of the SWCNTs. The unzipping and subsequent wrapping events are initiated and driven by van der Waals interactions between the aromatic rings of siRNA nucleobases and the SWCNT surface. However, molecular dynamics (MD) simulations of double strand DNA (dsDNA) of the same sequence show that the dsDNA undergoes much less unzipping and wrapping on the SWCNT in the simulation time scale of 70 ns. This interesting difference is due to smaller interaction energy of thymidine of dsDNA with the SWCNT compared to that of uridine of siRNA, as calculated by dispersion corrected density functional theory (DFT) methods. After the optimal binding of siRNA to SWCNT, the complex is very stable which serves as one of the major mechanisms of siRNA delivery for biomedical applications. Since siRNA has to undergo unwinding process with the effect of RNA-induced silencing complex, our proposed delivery mechanism by SWCNT possesses potential advantages in achieving RNA interference. (C) 2012 American Institute of Physics. [http://dx.doi.org/10.1063/1.3682780]

Fragment Finder 2.0: a computing server to identify structurally similar fragments

Fragment Finder 2.0 is a web-based interactive computing server which can be used to retrieve structurally similar protein fragments from 25 and 90% nonredundant data sets. The computing server identifies structurally similar fragments using the protein backbone C alpha angles. In addition, the identified fragments can be superimposed using either of the two structural superposition programs, STAMP and PROFIT, provided in the server. The freely available Java plug-in Jmol has been interfaced with the server for the visualization of the query and superposed fragments. The server is the updated version of a previously developed search engine and employs an in-house-developed fast pattern matching algorithm. This server can be accessed freely over the World Wide Web through the URL http://cluster.physics.iisc.ernet.in/ff/.

Carbon Nanotube based Composite Fibers for Strain Sensing, Signal Processing and Computing

Carbon nanotubes dispersed in polymer matrix have been aligned in the form of fibers and interconnects and cured electrically and by UV light. Conductivity and effective semiconductor tunneling against reverse to forward bias field have been designed to have differentiable current-voltage response of each of the fiber/channel. The current-voltage response is a function of the strain applied to the fibers along axial direction. Biaxial and shear strains are correlated by differentiating signals from the aligned fibers/channels. Using a small doping of magnetic nanoparticles in these composite fibers, magneto-resistance properties are realized which are strong enough to use the resulting magnetostriction as a state variable for signal processing and computing. Various basic analog signal processing tasks such as addition, convolution and filtering etc. can be performed. These preliminary study shows promising application of the concept in combined analog-digital computation in carbon nanotube based fibers. Various dynamic effects such as relaxation, electric field dependent nonlinearities and hysteresis on the output signals are studied using experimental data and analytical model.

Applying Genetic Algorithms to Optimize Power in Tiled SNUCA Chip Multicore Architectures

We propose a novel technique for reducing the power consumed by the on-chip cache in SNUCA chip multicore platform. This is achieved by what we call a "remap table", which maps accesses to the cache banks that are as close as possible to the cores, on which the processes are scheduled. With this technique, instead of using all the available cache, we use a portion of the cache and allocate lesser cache to the application. We formulate the problem as an energy-delay (ED) minimization problem and solve it offline using a scalable genetic algorithm approach. Our experiments show up to 40% of savings in the memory sub-system power consumption and 47% savings in energy-delay product (ED).

Computing Reeb Graphs as a Union of Contour Trees

The Reeb graph of a scalar function tracks the evolution of the topology of its level sets. This paper describes a fast algorithm to compute the Reeb graph of a piecewise-linear (PL) function defined over manifolds and non-manifolds. The key idea in the proposed approach is to maximally leverage the efficient contour tree algorithm to compute the Reeb graph. The algorithm proceeds by dividing the input into a set of subvolumes that have loop-free Reeb graphs using the join tree of the scalar function and computes the Reeb graph by combining the contour trees of all the subvolumes. Since the key ingredient of this method is a series of union-find operations, the algorithm is fast in practice. Experimental results demonstrate that it outperforms current generic algorithms by a factor of up to two orders of magnitude, and has a performance on par with algorithms that are catered to restricted classes of input. The algorithm also extends to handle large data that do not fit in memory.

A trust model for routing in MANETs: a cognitive agents based approach

Mobile ad hoc networks (MANETs) is one of the successful wireless network paradigms which offers unrestricted mobility without depending on any underlying infrastructure. MANETs have become an exciting and im- portant technology in recent years because of the rapid proliferation of variety of wireless devices, and increased use of ad hoc networks in various applications. Like any other networks, MANETs are also prone to variety of attacks majorly in routing side, most of the proposed secured routing solutions based on cryptography and authentication methods have greater overhead, which results in latency problems and resource crunch problems, especially in energy side. The successful working of these mechanisms also depends on secured key management involving a trusted third authority, which is generally difficult to implement in MANET environ-ment due to volatile topology. Designing a secured routing algorithm for MANETs which incorporates the notion of trust without maintaining any trusted third entity is an interesting research problem in recent years. This paper propose a new trust model based on cognitive reasoning,which associates the notion of trust with all the member nodes of MANETs using a novel Behaviors-Observations- Beliefs(BOB) model. These trust values are used for detec- tion and prevention of malicious and dishonest nodes while routing the data. The proposed trust model works with the DTM-DSR protocol, which involves computation of direct trust between any two nodes using cognitive knowledge. We have taken care of trust fading over time, rewards, and penalties while computing the trustworthiness of a node and also route. A simulator is developed for testing the proposed algorithm, the results of experiments shows incorporation of cognitive reasoning for computation of trust in routing effectively detects intrusions in MANET environment, and generates more reliable routes for secured routing of data.

A Hybrid Parallel Algorithm for Computing and Tracking Level Set Topology

The contour tree is a topological abstraction of a scalar field that captures evolution in level set connectivity. It is an effective representation for visual exploration and analysis of scientific data. We describe a work-efficient, output sensitive, and scalable parallel algorithm for computing the contour tree of a scalar field defined on a domain that is represented using either an unstructured mesh or a structured grid. A hybrid implementation of the algorithm using the GPU and multi-core CPU can compute the contour tree of an input containing 16 million vertices in less than ten seconds with a speedup factor of upto 13. Experiments based on an implementation in a multi-core CPU environment show near-linear speedup for large data sets.

Least squares QR-based decomposition provides an efficient way of computing optimal regularization parameter in photoacoustic tomography

A computationally efficient approach that computes the optimal regularization parameter for the Tikhonov-minimization scheme is developed for photoacoustic imaging. This approach is based on the least squares-QR decomposition which is a well-known dimensionality reduction technique for a large system of equations. It is shown that the proposed framework is effective in terms of quantitative and qualitative reconstructions of initial pressure distribution enabled via finding an optimal regularization parameter. The computational efficiency and performance of the proposed method are shown using a test case of numerical blood vessel phantom, where the initial pressure is exactly known for quantitative comparison. (C) 2013 Society of Photo-Optical Instrumentation Engineers (SPIE)

Hardness results for computing optimal locally Gabriel graphs

Delaunay and Gabriel graphs are widely studied geo-metric proximity structures. Motivated by applications in wireless routing, relaxed versions of these graphs known as Locally Delaunay Graphs (LDGs) and Lo-cally Gabriel Graphs (LGGs) have been proposed. We propose another generalization of LGGs called Gener-alized Locally Gabriel Graphs (GLGGs) in the context when certain edges are forbidden in the graph. Unlike a Gabriel Graph, there is no unique LGG or GLGG for a given point set because no edge is necessarily in-cluded or excluded. This property allows us to choose an LGG/GLGG that optimizes a parameter of interest in the graph. We show that computing an edge max-imum GLGG for a given problem instance is NP-hard and also APX-hard. We also show that computing an LGG on a given point set with dilation ≤k is NP-hard. Finally, we give an algorithm to verify whether a given geometric graph G= (V, E) is a valid LGG.

On Computing Amplitude, Phase, and Frequency Modulations Using a Vector Interpretation of the Analytic Signal

The amplitude-modulation (AM) and phase-modulation (PM) of an amplitude-modulated frequency-modulated (AM-FM) signal are defined as the modulus and phase angle, respectively, of the analytic signal (AS). The FM is defined as the derivative of the PM. However, this standard definition results in a PM with jump discontinuities in cases when the AM index exceeds unity, resulting in an FM that contains impulses. We propose a new approach to define smooth AM, PM, and FM for the AS, where the PM is computed as the solution to an optimization problem based on a vector interpretation of the AS. Our approach is directly linked to the fractional Hilbert transform (FrHT) and leads to an eigenvalue problem. The resulting PM and AM are shown to be smooth, and in particular, the AM turns out to be bipolar. We show an equivalence of the eigenvalue formulation to the square of the AS, and arrive at a simple method to compute the smooth PM. Some examples on synthesized and real signals are provided to validate the theoretical calculations.

MRT Letter: Light Sheet Based Imaging Flow Cytometry on a Microfluidic Platform

We propose a light sheet based imaging flow cytometry technique for simultaneous counting and imaging of cells on a microfluidic platform. Light sheet covers the entire microfluidic channel and thus omits the necessity of flow focusing and point scanning based technology. Another advantage lies in the orthogonal detection geometry that totally cuts-off the incident light, thereby substantially reducing the background in the detection. Compared to the existing state-of-art techniques the proposed technique shows marked improvement. Using fluorescently-coated Saccharomyces cerevisiae cells we have recorded cell counting with throughput as high as 2,090 cells/min in the low flow rate regime and were able to image the individual cells on-the-go. Overall, the proposed system is cost-effective and simple in channel geometry with the advantage of efficient counting in operational regime of low laminar flow. This technique may advance the emerging field of microfluidic based cytometry for applications in nanomedicine and point of care diagnostics. Microsc. Res. Tech. 76:1101-1107, 2013. (c) 2013 Wiley Periodicals, Inc.