Tom Kazmierski | University of Southampton (original) (raw)

Papers by Tom Kazmierski

Design, Automation, and Test in Europe, Apr 20, 2009

Dynamic thermal management is an increasingly critical and complex part of the run-time managemen... more Dynamic thermal management is an increasingly critical and complex part of the run-time management of manycore systems. Methods of controlling temperature include thread migration, dynamic voltage and frequency scaling and power gating using various strategies and combinations of each. In the PRiME project we are developing run-time management systems to sustain the scaling of many-core systems. As part of this development we are investigating the relative benefits of different thermal management strategies by co-simulating a Modellica model of the characteristics of a many-core device with a discrete Event-B model of the run-time manager. The results enable us to efficiently design more elaborate experiments on real hardware platforms in order to validate the run time management.

Lecture notes in electrical engineering, Dec 21, 2019

Lecture notes in electrical engineering, Nov 11, 2017

This paper provides the first analysis of process variation effect on the adiabatic logic combine... more This paper provides the first analysis of process variation effect on the adiabatic logic combined with nearthreshold operation. One of the significant concerns is whether reliable performance is retained. We find that typical variations of process parameters do not affect error-free operation at the minimum-energy frequency. Monte Carlo simulations of a 4- bit full adder using ECRL logic with 0.45 V supply voltage show that, in the presence of typical process variations, energy consumption of the circuit operating at 25 MHz increases by 10.2% in the worst case while a 100% error-free operation is maintained. The maximum operating frequency(208 MHz) is reduced to nearly half of the nominal value(385 MHz). To further improve the robustness of the adder against process variation, a bit-serial adiabatic adder is considered with an even lower energy consumption per cycle.

Lecture notes in electrical engineering, 2012

, except for brief excerpts in connection with reviews or scholarly analysis. Use in connection w... more , except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed is forbidden. The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights.

Design, Automation, and Test in Europe, Mar 8, 2010

Design, Automation, and Test in Europe, Mar 12, 2012

Three issues affect the reliability of circuit-level simulation. They are numerical stability of ... more Three issues affect the reliability of circuit-level simulation. They are numerical stability of integration, precision of results at each time step and convergence at each time step. The method underlying simulation is first the formulation of a coupled set of nonlinear first-order differential equations representing the behaviour of the interconnected set of devices comprising the circuit. Second, the replacement of the time derivatives in the differential equations by finite-difference approximations (known as integration formulae) which discretise time, in general, in a nonuniform way. This step transforms the nonlinear differential equations, at each discretised time point, into a time-independent set of nonlinear equations. The third step is to solve the nonlinear equations, at each discretised time, by the Newton-Raphson technique which approximates them with a linear set of equations based on an initial estimate of the solution. Repeated solution of the linear set, with appr...

Energy Harvesting Systems, 2010

ABSTRACT

2008 IEEE International Behavioral Modeling and Simulation Workshop, 2008

Sigma-Delta force-feedback control loop from user defined high-level performance specifications a... more Sigma-Delta force-feedback control loop from user defined high-level performance specifications and design constraints. The proposed approach is based on a simulationbased optimization technology using a genetic algorithm. The layout of the mechanical sensing element is generated simultaneously with the optimal design parameters of the Sigma-Delta control loop. As currently available implementations of AMS HDL languages are not suitable for complex mixed-technology system optimisation, the algorithm as well as aa fast dedicated sigma-delta accelerometer simulator have been implemented in C++. The underlying accelerometer model includes the sense finger dynamics described by a partial differential equation, which enables accurate performance prediction of the sensing element embedded in a in mixed-technology control loop.

Proceedings of 2010 IEEE International Symposium on Circuits and Systems, 2010

... of MEMS accelerometers with Sigma-Delta control Chenxu Zhao and Tom J. Kazmierski School of E... more ... of MEMS accelerometers with Sigma-Delta control Chenxu Zhao and Tom J. Kazmierski School of Electronics and Computer Science University of Southampton, UK Email: cz05r,tjk@ecs.soton. ac.uk ... dend1 dend2 Classic serpentine spring Rotated serpentine spring Proof Mass ...

Lecture Notes in Electrical Engineering, 2008

This paper proposes VHDL-AMS syntax extensions that enable descriptions of AMS systems with parti... more This paper proposes VHDL-AMS syntax extensions that enable descriptions of AMS systems with partial differential equations. We named the extended language VHDL-AMSP. An important specific need for such extensions arises from the well known MEMS modelling difficulties where complex digital and analogue electronics interfaces with distributed mechanical systems. The new syntax allows descriptions of new VHDL-AMS objects, such as partial quantities, spatial coordinates and boundary conditions. Pending the development of a new standard, a suitable preprocessor has been developed to convert VHDL-AMSP into the existing VHDL-AMS 1076.1 standard automatically. The preprocessor allows development of models with partial differential equations using currently available simulators. As an example, a VHDL-AMSP description for the sensing element of a MEMS accelerometer is presented, converted to VHDL-AMS 1076.1 and simulated in SystemVision.

2007 IEEE International Behavioral Modeling and Simulation Workshop, 2007

This contribution presents a novel MEMS accelerometer model implemented in VHDL-AMS. The model in... more This contribution presents a novel MEMS accelerometer model implemented in VHDL-AMS. The model includes sense finger dynamics, which allow accurate performance prediction of a MEMS accelerometer in a mixed-technology control loop. A distributed mechanical sensing element model is developed and the effect of the sense finger dynamics is analyzed. The sense finger dynamics might cause a failure of the Sigma-Delta control loop which is captured by the proposed model but cannot be correctly modeled using the conventional approach.

The 2010 IEEE International Conference on Information and Automation, 2010

This contribution presents a novel methodology for automated optimal design of MEMS accelerometer... more This contribution presents a novel methodology for automated optimal design of MEMS accelerometers embedded in electronic control circuitry from user deſned high-level performance speciſcations and design constraints. The proposed approach is based on simulation-based optimisation where automated conſguration selection for electronic blocks and synthesis of mechanical layouts are coupled with calculations of optimal design parameters. The underlying dedicated MEMS simulator supports distributed mechanical dynamics to enable accurate performance prediction of critical mechanical components, such as acceleration sensing elements which form an essential part of the mixed-technology control loop.

Lecture Notes in Electrical Engineering, 2011

Design, Automation, and Test in Europe, 2004

In this paper, a modified model of ferromagnetic hysteresis suitable for mixed-signal simulations... more In this paper, a modified model of ferromagnetic hysteresis suitable for mixed-signal simulations in VHDL-AMS is presented. The aim of this paper is to demonstrate how a numerically stable and accurate implementation of the Jiles-Atherton model can be achieved using a 4th order Runga-Kutta integration of the derivative of magnetization with respect to the field strength (H). While most SPICE-like

MILCOM 2008 - 2008 IEEE Military Communications Conference, 2008

Mismatches in the response of the components used in the analogue front-ends of adaptive beamform... more Mismatches in the response of the components used in the analogue front-ends of adaptive beamformers can reduce their ability to produce deep nulls against broadband interferers. These mismatches can be removed through a process of calibration performed at start-up and, if required, on a regular basis to track changes in the component's response due to environmental changes. In this paper a hardware efficient scheme for compensating for channel mismatches is proposed based on pre-equalization of each channel using frequency domain equalization. The effectiveness of the proposed scheme is simulated and the improvement in broadband nulling is demonstrated on a hardware implementation of an adaptive beamformer.