Lost in translation : Exposing hidden compiler optimization opportunities
- Submitting institution
-
University of Bristol
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 257474891
- Type
- D - Journal article
- DOI
-
10.1093/comjnl/bxaa103
- Title of journal
- The Computer Journal
- Article number
- bxaa103
- First page
- -
- Volume
- 0
- Issue
- -
- ISSN
- 0010-4620
- Open access status
- Compliant
- Month of publication
- August
- Year of publication
- 2020
- 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
-
4
- Research group(s)
-
D - Fundamentals of Computing
- Citation count
- -
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- This paper focuses on how optimization opportunities, that are not picked up by the state-of-the-art tools, can be detected and exploited to obtain better performance (i.e., faster, smaller, lower-energy code) and for debug. Results are at least as good if not better than those using Machine Learning, without requiring expensive training/retraining that ML needs - giving our technique a significant advantage, as commented by a reviewer “This is a key advance in compiler development with immediate practical application”. Already started attracting attention (e.g., invited talk at ISC High Performance 2020 "First Workshop on LLVM Compilers and Tools for High-Performance Computing").
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -