A methodology correlating code optimizations with data memory accesses, execution time and energy consumption
- Submitting institution
-
University of Plymouth
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 2066
- Type
- D - Journal article
- DOI
-
10.1007/s11227-019-02880-z
- Title of journal
- Journal of Supercomputing
- Article number
- -
- First page
- 6710
- Volume
- 75
- Issue
- 10
- ISSN
- 0920-8542
- Open access status
- Compliant
- Month of publication
- -
- Year of publication
- 2019
- 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
-
1
- Research group(s)
-
-
- Citation count
- 2
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- A novel methodology is delivered, that reduces the execution time and energy consumption of memory-bound loop kernels. For the first time, the execution time and power consumption of programs are expressed as mathematical functions of the source code optimization parameters and CPU hardware characteristics. Thereby, the efficient code optimization parameters can be found orders of magnitude faster than previously, allowing for vastly more efficient compilation of optimally efficient code. This work contributes to the TANGO software toolkit, which, for example, is exploited by the multi-national companies Bull (France) and Deltatec (Belgium) as part of their core service offerings (http://www.tango-project.eu/).
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -