Developing High Performance Computing Infrastructure and Research Protocols

Official Title
High Performance Computing for Large Scale Modeling and Simulation
Technology Solutions
Food and Drug Administration (FDA) - Center for Tobacco Product's Office of Science (CTP/OS)
High Performance Computing (HPC) architecture

Modern scientific research is increasingly dependent on large mathematical structures and datasets, which quickly stretch the limits of traditional desktop computing environments.

To help the U.S. Food and Drug Administration (FDA) assess the use of high performance computing (HPC) for processing large datasets, IMPAQ did the following:

  • Developed research protocols designed to guide researchers through a set of procedures to assess the desirability and practicality of applying HPC resources to large-scale modeling and simulation challenges.
  • Implemented a series of algorithms on high-performance cloud computing resources within the fields of Agent Based Modeling, System Dynamics Modeling, Social Network Analysis, and Text Analytics.
  • Produced technical reports to document and evaluate the performance of the various algorithms within an HPC setting.
  • Produced academic research articles using the experimental results.


The protocols developed by IMPAQ provided step-by-step guides to researchers wishing to implement an application in an HPC environment, including hardware considerations, cost, desirability and practicality. They included information on limitations posed by standard computing environments, opportunities offered by HPC resources to overcome computational problems, and trade-offs associated with using standard computing and HPC resources. The technical reports describe the parallelized algorithms, the implementation details, and the resulting increase in computational efficiency. The technical reports are also accompanied by the source code, which allows researchers to directly replicate our results.

Skills and Experience

For this project, the IMPAQ team consists of High Performance Computing (HPC) Engineers and Data Scientists with experience in optimization methods and data management.

In helping clients answer policy questions or manage resources, IMPAQ has leveraged datasets of different complexities through a wide range of tools and frameworks. This experience uniquely positions us to understand the limitations of standard computing environments, as well as the numerous opportunities offered by HPC resources. It also allows us to demonstrate our ability to dramatically scale-up existing algorithms.