Abstract
We investigate substituting a parallel, half-approximation for maximum weight matching instead of an exact computation. Our experiments show that the resulting difference in solution quality is negligible. We demonstrate almost a 20-fold speedup using 40 threads on an 8 processor Intel Xeon E7-8870 system and now solve real-world problems in 36 seconds instead of 10 minutes
- Developers:
-
Khan, Arif [1] ; Ghosh, Sayan [2]
- Pacific Northwest National Lab. (PNNL), Richland, WA (United States)
- Pacific Northwest National Laboratory
- Release Date:
- 2022-11-21
- Project Type:
- Open Source, Publicly Available Repository
- Software Type:
- Scientific
- Licenses:
-
BSD 2-clause "Simplified" License
- Sponsoring Org.:
-
USDOEPrimary Award/Contract Number:AC05-76RL01830
- Code ID:
- 97092
- Site Accession Number:
- Battelle IPID 32549-E
- Research Org.:
- Pacific Northwest National Laboratory (PNNL), Richland, WA (United States)
- Country of Origin:
- United States
Citation Formats
Khan, Arif, and Ghosh, Sayan.
ECP-ExaGraph/netAlign.
Computer Software.
https://github.com/ECP-ExaGraph/netAlign.
USDOE.
21 Nov. 2022.
Web.
doi:10.11578/dc.20221121.1.
Khan, Arif, & Ghosh, Sayan.
(2022, November 21).
ECP-ExaGraph/netAlign.
[Computer software].
https://github.com/ECP-ExaGraph/netAlign.
https://doi.org/10.11578/dc.20221121.1.
Khan, Arif, and Ghosh, Sayan.
"ECP-ExaGraph/netAlign." Computer software.
November 21, 2022.
https://github.com/ECP-ExaGraph/netAlign.
https://doi.org/10.11578/dc.20221121.1.
@misc{
doecode_97092,
title = {ECP-ExaGraph/netAlign},
author = {Khan, Arif and Ghosh, Sayan},
abstractNote = {We investigate substituting a parallel, half-approximation for maximum weight matching instead of an exact computation. Our experiments show that the resulting difference in solution quality is negligible. We demonstrate almost a 20-fold speedup using 40 threads on an 8 processor Intel Xeon E7-8870 system and now solve real-world problems in 36 seconds instead of 10 minutes},
doi = {10.11578/dc.20221121.1},
url = {https://doi.org/10.11578/dc.20221121.1},
howpublished = {[Computer Software] \url{https://doi.org/10.11578/dc.20221121.1}},
year = {2022},
month = {nov}
}