Adaptive Asynchronous Parallelization of Graph Algorithms
- Submitting institution
-
University of Edinburgh
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 162945433
- Type
- E - Conference contribution
- DOI
-
10.1145/3183713.3196918
- Title of conference / published proceedings
- Proceedings of the 2018 International Conference on Management of Data (SIGMOD'18)
- First page
- 1141
- Volume
- -
- Issue
- -
- ISSN
- 0730-8078
- Open access status
- Compliant
- Month of publication
- May
- Year of publication
- 2018
- 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
-
6
- Research group(s)
-
C - Foundations of Computation
- Citation count
- 9
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- Proposes an Adaptive Asynchronous Parallel (AAP) model for graph computations. As opposed to conventional Bulk Synchronous Parallel (BSP) and Asynchronous Parallel (AP) models, AAP reduces both stragglers and stale computations by dynamically adjusting relative progress of workers. BSP, AP and Stale Synchronous Parallel model are special cases of AAP. Better yet, AAP optimizes parallel processing by adaptively switching among these models at different stages of a single execution. The paper was one of only top-2 ranked SIGMOD papers invited to TODS 45 (https://dblp.org/db/journals/tods/tods45.html#FanLYXYLZJ20). AAP has been evaluated by Alibaba Group and proven effective in their real-life applications (contact: VP).
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -