Steal locally, share globally
- Submitting institution
-
University of Glasgow
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 11-02452
- Type
- D - Journal article
- DOI
-
10.1007/s10766-015-0350-0
- Title of journal
- International Journal of Parallel Programming
- Article number
- -
- First page
- 894
- Volume
- 43
- Issue
- 5
- ISSN
- 0885-7458
- Open access status
- Out of scope for open access requirements
- Month of publication
- October
- Year of publication
- 2015
- URL
-
http://eprints.gla.ac.uk/112246/
- 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
-
1
- Research group(s)
-
-
- Citation count
- 0
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- ORIGINALITY: This work presents a novel methodology for task-based shared-memory parallel programming for advanced manycore architectures. SIGNIFICANCE: Our methodology offers an intuitive task-based programming model where the run-time system takes care of the speed-up - which is important for reducing errors in critical applications. Moreover, it markedly improves the performance of applications where multiple parallel workloads compete for system resources. Published in one of the 3 top parallel programming journals. RIGOUR: Using a state-of-the-art manycore system, we validated our methodology against three well-established parallel programming models using three benchmarks each stress-testing different system performance aspects and two complex multiprogramming scenarios.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -