Adaptive Scalable SVD Unit for Fast Processing of Large LSE Problems

Singular Value Decomposition (SVD) is a key linear algebraic operation in many scientific and engineering applications. In particular, many computational intelligence systems rely on machine learning methods involving high dimensionality datasets that have to be fast processed for real-time adaptability. In this paper we describe a practical FPGA (Field Programmable Gate Array) implementation of a SVD processor for accelerating the solution of large LSE problems. The design approach has been comprehensive, from the algorithmic refinement to the numerical analysis to the customization for an efficient hardware realization. The processing scheme rests on an adaptive vector rotation evaluator for error regularization that enhances convergence speed with no penalty on the solution accuracy. The proposed architecture, which follows a data transfer scheme, is scalable and based on the interconnection of simple rotations units, which allows for a trade-off between occupied area and processing acceleration in the final implementation. This permits the SVD processor to be implemented both on low-cost and highend FPGAs, according to the final application requirements.

Scalable Methods for Deterministic Integration of Quantum Emitters in Photonic Crystal Cavities

We investigated four unique methods for achieving scalable, deterministic integration of quantum emitters into ultra-high Q{V photonic crystal cavities, including selective area heteroepitaxy, engineered photoemission from silicon nanostructures, wafer bonding and dimensional reduction of III-V quantum wells, and cavity-enhanced optical trapping. In these areas, we were able to demonstrate site-selective heteroepitaxy, size-tunable photoluminescence from silicon nanostructures, Purcell modification of QW emission spectra, and limits of cavity-enhanced optical trapping designs which exceed any reports in the literature and suggest the feasibility of capturing- and detecting nanostructures with dimensions below 10 nm. In addition to process scalability and the requirement for achieving accurate spectral- and spatial overlap between the emitter and cavity, these techniques paid specific attention to the ability to separate the cavity and emitter material systems in order to allow optimal selection of these independently, and eventually enable monolithic integration with other photonic and electronic circuitry.

We also developed an analytic photonic crystal design process yielding optimized cavity tapers with minimal computational effort, and reported on a general cavity modification which exhibits improved fabrication tolerance by relying exclusively on positional- rather than dimensional tapering. We compared several experimental coupling techniques for device characterization. Significant efforts were devoted to optimizing cavity fabrication, including the use of atomic layer deposition to improve surface quality, exploration into factors affecting the design fracturing, and automated analysis of SEM images. Using optimized fabrication procedures, we experimentally demonstrated 1D photonic crystal nanobeam cavities exhibiting the highest Q/V reported on substrate. Finally, we analyzed the bistable behavior of the devices to quantify the nonlinear optical response of our cavities.

Scalable quantum memory in the ultrastrong coupling regime

Circuit quantum electrodynamics, consisting of superconducting artificial atoms coupled to on-chip resonators, represents a prime candidate to implement the scalable quantum computing architecture because of the presence of good tunability and controllability. Furthermore, recent advances have pushed the technology towards the ultrastrong coupling regime of light-matter interaction, where the qubit-resonator coupling strength reaches a considerable fraction of the resonator frequency. Here, we propose a qubit-resonator system operating in that regime, as a quantum memory device and study the storage and retrieval of quantum information in and from the Z(2) parity-protected quantum memory, within experimentally feasible schemes. We are also convinced that our proposal might pave a way to realize a scalable quantum random-access memory due to its fast storage and readout performances.

A novel scalable photonic analog switch architecture based on semiconductor optical amplifiers

We propose a novel semiconductor optical amplifier (SOA) based switch architecture for analog applications. Proof-of-principle experiments show that the system is very linear with an SFDR of approximately 100dB·Hz 2/3 for a switching time of 50μs. The port number of this switch is scalable and can be expanded to 80 × 80.

Low-cost, scalable optical packet switching networks with multi-wavelength labels

We propose a self-forwarding packet-switched optical network with bit-parallel multi-wavelength labels. We experimentally demonstrate transmission of variable-length optical packets over 80 km of fiber and switching over a 1×4 multistage switch with two stages. © 2007 Optical Society of America.

A practical demonstration of scalable, parallel mesh generation

Cambridge Flow Solutions Ltd, Compass House, Vision Park, Cambridge, CB4 9AD, UK Real-world simulation challenges are getting bigger: virtual aero-engines with multistage blade rows coupled with their secondary air systems & with fully featured geometry; environmental flows at meta-scales over resolved cities; synthetic battlefields. It is clear that the future of simulation is scalable, end-to-end parallelism. To address these challenges we have reported in a sequence of papers a series of inherently parallel building blocks based on the integration of a Level Set based geometry kernel with an octree-based cut-Cartesian mesh generator, RANS flow solver, post-processing and geometry management & editing. The cut-cells which characterize the approach are eliminated by exporting a body-conformal mesh driven by the underpinning Level Set and managed by mesh quality optimization algorithms; this permits third party flow solvers to be deployed. This paper continues this sequence by reporting & demonstrating two main novelties: variable depth volume mesh refinement enabling variable surface mesh refinement and a radical rework of the mesh generation into a bottom-up system based on Space Filling Curves. Also reported are the associated extensions to body-conformal mesh export. Everything is implemented in a scalable, parallel manner. As a practical demonstration, meshes of guaranteed quality are generated for a fully resolved, generic aircraft carrier geometry, a cooled disc brake assembly and a B747 in landing configuration. Copyright © 2009 by W.N.Dawes.

Using Level Sets as the basis for a scalable, parallel geometry engine and mesh generation system

The background to this review paper is research we have performed over recent years aimed at developing a simulation system capable of handling large scale, real world applications implemented in an end-to-end parallel, scalable manner. The particular focus of this paper is the use of a Level Set solid modeling geometry kernel within this parallel framework to enable automated design optimization without topological restrictions and on geometries of arbitrary complexity. Also described is another interesting application of Level Sets: their use in guiding the export of a body-conformal mesh from our basic cut-Cartesian background octree - mesh - this permits third party flow solvers to be deployed. As a practical demonstrations meshes of guaranteed quality are generated and flow-solved for a B747 in full landing configuration and an automated optimization is performed on a cooled turbine tip geometry. Copyright © 2009 by W.N.Dawes.

Scalable robustness for consensus protocols with heterogeneous dynamics

It is shown in the paper how robustness can be guaranteed for consensus protocols with heterogeneous dynamics in a scalable and decentralized way i.e. by each agent satisfying a test that does not require knowledge of the entire network. Random graph examples illustrate that the proposed certificates are not conservative for classes of large scale networks, despite the heterogeneity of the dynamics, which is a distinctive feature of this work. The conditions hold for symmetric protocols and more conservative stability conditions are given for general nonsymmetric interconnections. Nonlinear extensions in an IQC framework are finally discussed. Copyright © 2005 IFAC.