21 October 2014 Optimizing zonal advection of the Advanced Research WRF (ARW) dynamics for Intel MIC
Author Affiliations +
Abstract
The Weather Research and Forecast (WRF) model is the most widely used community weather forecast and research model in the world. There are two distinct varieties of WRF. The Advanced Research WRF (ARW) is an experimental, advanced research version featuring very high resolution. The WRF Nonhydrostatic Mesoscale Model (WRF-NMM) has been designed for forecasting operations. WRF consists of dynamics code and several physics modules. The WRF-ARW core is based on an Eulerian solver for the fully compressible nonhydrostatic equations. In the paper, we will use Intel Intel Many Integrated Core (MIC) architecture to substantially increase the performance of a zonal advection subroutine for optimization. It is of the most time consuming routines in the ARW dynamics core. Advection advances the explicit perturbation horizontal momentum equations by adding in the large-timestep tendency along with the small timestep pressure gradient tendency. We will describe the challenges we met during the development of a high-speed dynamics code subroutine for MIC architecture. Furthermore, lessons learned from the code optimization process will be discussed. The results show that the optimizations improved performance of the original code on Xeon Phi 5110P by a factor of 2.4x.
© (2014) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Jarno Mielikainen, Bormin Huang, Allen H.-L. Huang, "Optimizing zonal advection of the Advanced Research WRF (ARW) dynamics for Intel MIC", Proc. SPIE 9247, High-Performance Computing in Remote Sensing IV, 92470M (21 October 2014); doi: 10.1117/12.2069314; https://doi.org/10.1117/12.2069314
PROCEEDINGS
11 PAGES


SHARE
Back to Top