Postgraduate Studentships
Proposed Project Title: Exploiting Application Error-Resilience for Designing Adaptive and Energy-Efficient Systems
Principal Supervisor: Dr. Georgios Karakonstantis

Project Description:

Currently, manufacturers go to great lengths to guardband the designed hardware against any potential functional failure induced by the highly unreliable nanometer silicon and the applied scaled voltages. Existing measures may have accomplished to hide any inaccurate hardware behaviour from the software layers and maintain acceptable yield levels, but unfortunately the large energy, performance, and area overheads that they incur limit their viability, especially as we move beyond the 45nm node. The incurred overheads have made the research community to look for alternative computing paradigms in which inaccurate computation and storage could be accepted. Recent experiments have shown the potential of such a paradigm shift in saving considerable amounts of energy in various applications such as multimedia and data mining the statistical nature of which can help in tolerating a number of inaccurate operations. However, there is still a need to explore the inherent ‘error’ resilience of many other applications and capture systematically the interplay between energy, reliability, and quality at different layers of design abstraction.

The proposed project plans to exploit the inherent error resilience and certain statistical characteristics of various applications for achieving adequately-reliable and energy efficient operation, while limiting or even avoiding the penalties incurred by traditional approaches. This will be achieved by developing low cost mechanisms at the architecture layer while considering the implications and needs at the circuit and software layer. The developed mechanisms will allow the adaptation of any system to dynamically changing operating conditions and user requirements and is expected to be beneficial for a variety of application domains from embedded to high-performance systems.

The 3-year PhD studentship will be based at Queen's University Belfast in the High Performance and Distributed Computing (HPDC) cluster with close collaboration with other top universities such as EPFL, Switzerland and industrial labs in U.K. and Europe.

Objectives:

• Develop a simulation framework for exploring the error resilient nature of various applications and analysing their statistical properties. Popular algorithms will be targeted and analysed from a variety of application domains including multimedia, biomedical as well as data mining and classification.
• Based on the identified properties develop mechanisms that could ensure good-enough output quality even under large number of functional failures. The developed mechanisms will target logic, arithmetic or memory sub-modules of a many-core system.
• Map the developed mechanisms on FPGA based systems and evaluate their efficacy.

Academic Requirements:

A minimum 2.1 honours degree or equivalent in Computer Science or Electrical and Electronic Engineering or relevant degree is required.

GENERAL INFORMATION

This 3 year PhD studentship, potentially funded by the Department for Employment and Learning (DEL), commences on 1 October 2015.

Eligibility for both fees and maintenance (£13,863 in 2014/15, 2015/16 TBC) depends on the applicants being either an ordinary UK resident or those EU residents who have lived permanently in the UK for the 3 years immediately preceding the start of the studentship. Non UK residents who hold EU residency may also apply but if successful may receive fees only.

Please note: DEL awards are available for Home and EU candidates only.

Applicants should apply electronically through the Queen’s online application portal at:https://dap.qub.ac.uk/portal/

Further information available at: http://www.qub.ac.uk/schools/eeecs/StudyattheSchool/PhDProgrammes

Contact detail

Supervisor Name: Dr. Georgios Karakonstanti
Email: georgios.karakonstantis<στο>epfl.ch, G.Karakonstantis<στο>qub.ac.uk

Deadline for submission of applications is 27 February 2015

For further information on Research Area click on link below:
http://www.qub.ac.uk/research-centres/HPDC/