Large-scale automatic k-means clustering for heterogeneous many-core supercomputer
- Submitting institution
-
University of St Andrews
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 264077482
- Type
- D - Journal article
- DOI
-
10.1109/TPDS.2019.2955467
- Title of journal
- IEEE Transactions on Parallel and Distributed Systems
- Article number
- -
- First page
- 997
- Volume
- 31
- Issue
- 5
- ISSN
- 1045-9219
- Open access status
- Compliant
- Month of publication
- November
- 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
-
8
- Research group(s)
-
B - Systems
- Citation count
- 0
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- k-means is a well known and heavily used algorithm in a number of domains. This work presented a new way of memory organisation and data partitioning for heterogeneous multi-level supercomputers, and is demonstrated on what was the world’s leading supercomputer at the time, the Sunway Taihulight. This allows k-means to operate with a far larger number of centroids and dimensions than was achievable before, and thus greater precision in deployment in applications, as demonstrated in the paper by object recognition on satellite imaging.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -
