| EPSRC Reference: |
GR/J91883/01 |
| Title: |
HIGH-LEVEL SHARING FOR PARALLEL PROGRAMMING |
| Principal Investigator: |
Dew, Professor P |
| Other Investigators: |
|
| Researcher Co-Investigators: |
|
| Project Partners: |
|
| Department: |
Sch of Computing |
| Organisation: |
University of Leeds |
| Scheme: |
Standard Research (Pre-FEC) |
| Starts: |
05 June 1994 |
Ends: |
04 July 1997 |
Value (£): |
125,551
|
| EPSRC Research Topic Classifications: |
|
| EPSRC Industrial Sector Classifications: |
|
| Related Grants: |
|
| Panel History: |
|
|
Summary on Grant Application Form |
|
(1) To identify a set of high-level abstractions for sharing and investigate their application in parallel programming (2) To demonstrate scalable and efficient implementation on both shared memory and distributed memory MIMD systems and on distributed networks of workstations (3) To investigate performance-improving transformations for selected abstractions, exploiting the characteristics of target machines.Progress:The TallShiP is a collaborative project between the University of Leeds and Rutherford Appleton Laboratory which addresses the fundamental role of sharing in parallel and distributed systems. The project is investigating sharing using typed data abstractions called Shared Abstract Data-types (SADTs) which should provide a suitable base-level for programming concurrent applications.The project is studying a number of practical issues. The key one is the consistency semantics. To construct programs that behave correctly with respect to sequential reasoning about their behaviour, all data must be sequentially consistent. This has lead on to study the linearisability condition of Herlihy and Wing which offers a uniform approach to consistency for all SADT implementations. A methodology for the verification and validation of linearised SADTs is being developed together with a simple transformational method. The project is using a number of case studies to clarify ideas and develop the methodology. The work at Leeds has focused on the implementation of highly concurrent linearisable SADTs and their use in building parallel applications. A three-stage methodology is being studied: an executable sequential specification in Standard ML is refined to a reference implementation in Concurrent ML. This is then translated into a target implementation for a specific architecture. Two target implementations have been used so far - WPRAM (developed under grant GR/H77811 and MPI. A case study based on a high distributed concurrent FIFO queue has been used to explore this methodology with considerable success. Further case studies are planned based on a shared accumulator. The project will continue to explore the development of primitive set of SADTs which can be used to build real programs and which will support a theory of sharing being developed by Chris Wadworth and Simon Dobson at Rutherford Appleton Laboratory.
|
|
Key Findings |
|
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
|
|
Potential use in non-academic contexts |
|
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
|
|
Impacts |
|
| Description |
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk |
| Summary |
|
| Date Materialised |
|
|
|
|
Sectors submitted by the Researcher |
|
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
|
| Project URL: |
|
| Further Information: |
|
| Organisation Website: |
http://www.leeds.ac.uk |