Enabling efficient execution of a variational data assimilation application

Abstract

Remote sensing observational instruments are critical for better understanding and predicting severe weather. Observational data from such instruments, such as Doppler radar data, for example, are often processed for assimilation into numerical weather prediction models. As such instruments become more sophisticated, the amount of data to be processed grows and requires efficient variational analysis tools. Here we examine the code that implements the popular SAMURAI (Spline Analysis at Mesoscale Utilizing Radar and Aircraft Instrumentation) technique for estimating the atmospheric state for a given set of observations. We employ a number of techniques to significantly improve the code’s performance, including porting it to run on standard HPC clusters, analyzing and optimizing its single-node performance, implementing a more efficient nonlinear optimization method, and enabling the use of GPUs via OpenACC. Our efforts thus far have yielded more than 100x improvement over the original code on large test problems of interest to the community.

Variational Data Assimilation Method Using Parallel Dual Populations Particle Swarm Optimization Algorithm

“…g) The grouping thread calculates the relevant parameters according to formulae (8), ( 9), (10) and deter-mines whether each exchanged difference vector can replace the optimal particle of another population.…”

“…To address the observational data from remote sensing instruments, Dennis et al [10] scrutinized the code that operationalizes the widely used Spline Analysis at Mesoscale Utilizing Radar and Aircraft Instrumentation (SAMURAI) technique for estimating atmospheric conditions based on a designated collection of observations. They deployed several strategies to substantially enhance the code􀆳s efficiency, encompassing adapting it for operation on typical high-performance computing (HPC) clusters, evaluating and refining its single-node performance, introducing a more efficient nonlinear optimization approach, and facilitating Graphics Processing Unit (GPU) utilization through Open Accelerators (OpenACC).…”