Document Type
Conference Proceeding
Publication Date
11-13-2014
Publication Title
Proceedings of the International Conference on Parallel Processing
Abstract
We present the progression of developments necessary to scale the ISAM landsurface model from single nodes and small clusters with unusually largeper-node memory to much larger systems with more common configurations. These efforts include load balancing, conventional library-based output parallelization to reduce memory load, and parallel-in-time data input. OnHopper, a Cray XE6 machine, the result was strong scaling from 256 cores to 16k coreswith an efficiency of 32.9%. On Edison, a Cray XC30 machine, thecode strong scales from 256 cores to 16k coreswith an efficiency of 51.4%. These large-scale gains, and the associated performance increases at smaller scale, enable greater scientific productivity for the users of ISAM and open the possibilities of increased resolution in time and space and greater physical fidelity for the simulated processes while remaining computationally feasible.
Volume
2014-November
Issue
November
First Page
422
Last Page
431
DOI
10.1109/ICPP.2014.51
ISSN
01903918
Recommended Citation
Miller, Phil; Robson, Michael; El-Masri, Bassil; Barman, Rahul; Zheng, Gengbin; Jain, Atul; and Kalé, Laxmikant, "Scaling the ISAM land surface model through parallelization of inter-component data transfer" (2014). Computer Science: Faculty Publications, Smith College, Northampton, MA.
https://scholarworks.smith.edu/csc_facpubs/357
Comments
Archived as published.