Performance optimization of multi-core grammatical evolution generated parallel recursive programs
- Submitting institution
-
Birmingham City University
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 11Z_OP_E0012
- Type
- E - Conference contribution
- DOI
-
10.1145/2739480.2754746
- Title of conference / published proceedings
- GECCO '15: Proceedings of the Genetic and Evolutionary Computation Conference 2016
- First page
- 1007
- Volume
- -
- Issue
- -
- ISSN
- -
- Open access status
- -
- Month of publication
- -
- Year of publication
- 2015
- URL
-
https://dl.acm.org/doi/10.1145/2739480.2754746
- Supplementary information
-
-
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- No
- Number of additional authors
-
-
- Research group(s)
-
-
- Citation count
- 9
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- This is the first instance of using AI techniques like Evolutionary Algorithms to optimise the degree of parallelism and simultaneously produce parallel programs.
Writing parallel programs is hard, and automatically doing so is much harder. The challenge doubles up by optimising the number of threads. This work does all that and outperforms the best human produced results, winning the silver HUMIES award. (12th award) http://www.human-competitive.org/awards
The work tackles many challenging problems such as sorting, bases on well-researched principles of automatically producing programs and justifies performance gains with statistical rigour.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -