| EPSRC Reference: |
GR/H46893/01 |
| Title: |
AN INVESTIGATION OF A DISTRIBUTED PARALLEL ASSOCIATIVE PROCESSOR FOR THE EXECUTION OF LOGIC PROGRAMS |
| Principal Investigator: |
Glover, Dr R |
| Other Investigators: |
|
| Researcher Co-Investigators: |
|
| Project Partners: |
|
| Department: |
Electronic & Computer Engineering |
| Organisation: |
Brunel University London |
| Scheme: |
Standard Research (Pre-FEC) |
| Starts: |
01 October 1992 |
Ends: |
30 September 1995 |
Value (£): |
84,382
|
| EPSRC Research Topic Classifications: |
|
| EPSRC Industrial Sector Classifications: |
|
| Related Grants: |
|
| Panel History: |
|
|
Summary on Grant Application Form |
|
To investigate the implementation of a Prolog compiler for the ASTRA massively parallel associative processor made by Aspex Microsystems Ltd. and Brunel University. To benchmark the execution of logic programs on ASTRA and identify suitable applications where the parallelism of the machine may be exploited.Progress:ASTRA is essentially a Sun workstation host with a 16,384 processor single instruction multiple data stream associative processor closely coupled via an intermediate controller. Each processor has a 64 bit word of local content addressable memory. Architectural investigations identified three areas where ASTRA had potential to accelerate the execution of logic programs: data parallelism, clause filtering and stack management.In order to exploit data parallelism fully it was considered necessary to extend Prolog to include the set data type. Initial investigations centred on devising a suitable associative data structure to represent Prolog data entities, including list structures and the newly incorporated set type. Mapping this data structure on to the ASTRA architecture has highlighted the architectural parameters that need to be considered in the design of an associative processor for the execution of logic programs. A reasonable compromise has been obtained for the ASTRA implementation which is sub-optimal for Prolog because ASTRA is optimised for image processing applications.The use of an appropriate data representation and the associative manipulation of the internal Prolog execution stacks has shown sequential operation to be beneficially accelerated such that execution times for the n-queens benchmark matches those of optimised Prolog machines. The use of associative look-up for clause filtering is not considered to be important in static Prolog programs where compiler technology can match the advantages of the parallelism given by the associative processor. However, the use of Prolog to provide intelligent manipulation of large databases is an area identified where associatively accelerated interpreted Prolog may find wider application and a Prolog-SQL translator has been developed to allow ongoing investigation of database applications to proceed. Further investigations have considered parallel heuristic search strategies for Prolog which could be implemented by a LAN cluster of ASTRA workstations and is complementary with research elsewhere on parallel AND/OR execution.
|
|
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.brunel.ac.uk |