Enhance GPU/CPU Hybrid Parallel Computation for Parachute Simulation with Machine Learning
Abstract
We request funding to upgrade our parallel cluster as a computational testbed for the numerical study of fluid-interface interaction problems. We propose to build a transitional platform for the development of the parachute code for super-computing on the DOD HPC facilities. This proposal is an amendment to the grant W911NF-18-1-0346 for ÒAn Efficient Computational Platform for Simulation of Parachute and Other Deformable Structures in Turbulent FlowÓ The new request includes adding 20 Pinnacle Computer Block Servers to the existing cluster which was purchased through the 2015 DURIP grant W911NF-15-1-0403. This purchase will expand our current cluster to 620 cores and enable us to carry out medium sized simulation on the problems involving fluid and structure interactions. The proposed equipment will also give us the capability of testing computation intensive machine learning algorithms which can be applied to several components in the front tracking method and its application to the parachute simulation platform.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jul 09, 2020
- Source ID
- W911NF2010159
Entities
People
- Xiaolin Li
Organizations
- Army Contracting Command
- Research Foundation for the State University of New York
- United States Army