The University of Birmingham

Digital Systems and Vision Processing



About

People

Research

Projects

Resources

Opportunities

Publications


PhD Proposals

Supervisor: Dr S F Quigley

Computer architectures

Many important computational tasks require the performance of the same sequence of simple operations on very large datasets, to produce answers in real time. Examples include speech, image and signal processing, computer graphics and data compression. These tasks are often handled poorly by conventional computers.
Modification to the computer hardware can produce an architecture that is more suitable for performing massively parallel execution of simple instructions. Two examples are

  • reconfigurable computing, in which the computer contains an array of reconfigurable logic, that continuously re-organises itself to form a coprocessor to accelerate the task in hand.
  • active memories, in which an array of simple processors is embedded within main memory, graphics memory, network controllers, etc.
These approaches have the potential to give speed ups of several orders of magnitude. However, they create difficult challenges for the design of both the hardware and the operating system. They are also very sensitive to how efficiently the algorithm is mapped onto the hardware. PhD projects are available for investigating research issues in hardware, systems software or algorithm mapping for reconfigurable computers or active memory computers.

Back to Opportunities

 

Behavioural Synthesis of Analogue Integrated Circuits

For digital ICs, design methodologies exist that allow the desired function of a system to be specified (usually through a hardware description language), and to have this behaviour automatically synthesized to silicon. Not only does this make the design process easier and faster, it also facilitates rapid re-targeting of a design from one fabrication to another.

The aim of this project is to investigate approaches to providing a similar path to silicon for analogue circuits, entailing investigation of the issues involved in:

  • Design capture
  • Circuit-level implementation
  • Silicon-level implementation
  • Synthesis procedures
  • Back-annotation and post-layout simulation
  • Impact of operating conditions (e.g. frequency, temperature)

Under each of these headings, issues will be identified that promote or inhibit stable, predictable and reproducible behaviour. The path to silicon that is most promising will be identified and trialled for a range of typical circuits.

The successful applicant will work in conjunction with TRW Aeronautical Systems in Birmingham, and will receive £8000 pa in year one, rising to £9000 pa in the third year. International travel may also be possible, allowing time to be spent at one of TRW’s sites outside of the UK. It is expected that the successful applicant will have a degree within Electronic and Electrical Engineering – preferably a 1st or a high 2:1 -, however applicants from a Physics or Computer Science background will also be considered. Please note that due to the nature of the business that TRW operates in, and the fact that the work will be highly classified, it is regretted that applications cannot be accepted from non-UK citizens.

Back to Opportunities


School of EEE University Last updated by Jonathan Mangnall on 25th May 2001.