Parallelizing a Computationally Intensive Financial R Application with Zircon Technology Zircon Computing LLC

Statisticians, analysts, scientists, and engineers requi massive processing power to conduct data analysis, predictive modeling, visualization, and other complex tasks. Thi s paper describes how we substantially improved the performance of a representative complex computational finance application by integrating the Zircon adaptive ultra highperformance computing software platform and tools with the R programming language and environment. This integrated solution uses distribution and parallelization to r educe the total computation time of the R-based application from 3,093 minutes to 40 minutes on a commodity multiprocessing platform.