Automatic Optimization of Thread-coarsening for Graphics Processors
- Submitting institution
-
University of Edinburgh
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 58556543
- Type
- E - Conference contribution
- DOI
-
10.1145/2628071.2628087
- Title of conference / published proceedings
- PACT '14: Proceedings of the 23rd international conference on Parallel architectures and compilation
- First page
- 455
- Volume
- -
- Issue
- -
- ISSN
- 1089-795X
- Open access status
- Out of scope for open access requirements
- Month of publication
- August
- Year of publication
- 2014
- URL
-
-
- 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
-
2
- Research group(s)
-
A - Computer Systems
- Citation count
- 22
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- This paper is the first to use a machine learning based approach to automatically determine thread coarsening without any program prior knowledge or profile runs. Despite the problem being open since 2000 and its importance to performance, no other paper was able to solve this problem automatically. It was published at the top-tier conference PACT and is the most widely cited paper in the area. It's public data set has allowed take up in the wider ML and programming community and has become a standard evaluation task. This work led to close GPU collaboration with ARM.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -