A Deep-Learning Approach Towards Auto-Tuning CFD Codes

• funded by: AFOSR
• funding level: $90,000 (for NCSU)
• duration: 06/01/2017 - 02/14/2018
• PIs (total funding: $250,000):
  • Wuchun Feng, Danesh Tafti, Chris Roy, Eric de Sturler, and Adrian Sandu - Virginia Tech
  • Frank Mueller, Jack Edwards, and Hong Luo - North Carolina State University

This proposal seeks to study, analyze, and synthesis deep-learning approaches for heterogeneous computing devices that expose the various parameters as "knobs" that can be tuned via deep learning to optimize for the metric of interest, whether it be performance, power, or energy efficiency.

• "FuncyTuner: Auto-tuning Scientific Applications With Per-loop Compilation" by Tao Wang, Nikhil Jain, David Beckingsale, David Boehme, Frank Mueller, Todd Gamblin in International Conference on Parallel Processing (ICPP), Aug 2019, Best Paper Candidate.

• "Compiler-based Auto-tuning and Synchronization Validation for HPC Applications" by T. Wang, Ph.D. Thesis, North Carolina State University, Dec 2019 (last known position: TBD)