Deep and Shallow convections in Atmosphere Models on Intel (R) Xeon Phi (TM) Coprocessor Systems

Ramesh, Srinivasan and Vadhiyar, Sathish and Nanjundiah, Ravi and Vinayachandran, PN (2018) Deep and Shallow convections in Atmosphere Models on Intel (R) Xeon Phi (TM) Coprocessor Systems. In: 13th International Manufacturing Science and Engineering Conference, 18-20 Dec. 2017, Bangkok, Thailand, pp. 316-323.

PDF 2017_IEEE_19th _int_con_316-323_2018.pdf - Published Version Restricted to Registered users only Download (288kB) | Request a copy

Abstract

Deep and shallow convection calculations occupy significant times in atmosphere models. These calculations also present significant load imbalances due to varying cloud covers over different regions of the grid. In this work, we accelerate these calculations on Intel (R) Xeon Phi (TM) Coprocessor Systems. By employing dynamic scheduling in OpenMP, we demonstrate large reductions in load imbalance and about 10% increase in speedups. By careful categorization of data as private, firstprivate and shared, we minimize data copying overheads for the coprocessors. We identify regions of false sharing among threads and eliminate them by loop rearrangements. We also employ proportional partitioning of independent column computations across both the CPU and coprocessor cores based on the performance ratio of the computations on the heterogeneous resources. These techniques along with various vectorization strategies resulted in about 30% improvement in convection calculations.

Item Type:	Conference Proceedings
Publisher:	IEEE
Additional Information:	Copyright for this paper belongs to IEEE.
Keywords:	Intel Xeon Phi co-processors;offloading;CAM;convections
Department/Centre:	Division of Interdisciplinary Sciences > Computational and Data Sciences Division of Mechanical Sciences > Centre for Atmospheric & Oceanic Sciences
Date Deposited:	31 Dec 2018 10:52
Last Modified:	31 Dec 2018 10:52
URI:	http://eprints.iisc.ac.in/id/eprint/61268

Actions (login required)

View Item