Exploiting the chaotic behaviour of atmospheric models with reconfigurable architectures
- Submitting institution
-
Imperial College of Science, Technology and Medicine
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 2405
- Type
- D - Journal article
- DOI
-
10.1016/j.cpc.2017.08.011
- Title of journal
- Computer Physics Communications
- Article number
- -
- First page
- 160
- Volume
- 221
- Issue
- 12
- ISSN
- 0010-4655
- Open access status
- Compliant
- Month of publication
- August
- Year of publication
- 2017
- URL
-
-
- Supplementary information
-
10.1016/j.cpc.2017.08.011
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- Yes
- Number of additional authors
-
4
- Research group(s)
-
-
- Citation count
- 6
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- This novel reduced-precision optimisation method for reconfigurable architectures exploits the chaotic behaviour of atmospheric models to achieve more effective resource utilisation. It enables a new generation of weather and climate simulators, targeting such architectures in datacentres, which have the potential to compete with costly supercomputers. The conference version in FCCM’15 (https://doi.org/10.1109/FCCM.2015.52; acceptance rate: 22%/95) was nominated for a Best Paper award. Underpinned an EPSRC Platform Grant (EP/P010040/1; £1.2M) and led to a new collaboration with the European Centre for Medium-Range Weather Forecasts and Maxeler Technologies on exploring exascale computing for climate applications (Maxeler, contact: FoEREF@ic.ac.uk).
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -
