Bayesian Optimization for auto-tuning GPU kernels

• Published in: 2021 International Workshop on Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems (PMBS) pp. 106 - 117
• Main Authors: Willemsen, Floris-Jan, van Nieuwpoort, Rob, van Werkhoven, Ben
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.11.2021
• Subjects: auto-tuning; Bayes methods; Bayesian Optimization; Computational modeling; Convolution; GPU Computing; Graphics processing units; machine learning; Optimization; Scalability; Search problems; Tuners
• DOI: 10.1109/PMBS54543.2021.00017

Finding optimal parameter configurations for tunable GPU kernels is a non-trivial exercise for large search spaces, even when automated. This poses an optimization task on a nonconvex search space, using an expensive to evaluate function with unknown derivative. These characteristics make a good candidate for Bayesian Optimization, which has not been applied to this problem before. However, the application of Bayesian Optimization to this problem is challenging. We demonstrate how to deal with the rough, discrete, constrained search spaces, containing invalid configurations. We introduce a novel contextual variance exploration factor, as well as new acquisition functions with improved scalability, combined with an informed acquisition function selection mechanism. By comparing the performance of our Bayesian Optimization implementation on various test cases to the existing search strategies in Kernel Tuner, as well as other Bayesian Optimization implementations, we demonstrate that our search strategies generalize well and consistently outperform other search strategies by a wide margin.