Generating GPU Compiler Heuristics using Reinforcement Learning
This addresses the challenge of manual heuristic design in GPU compilers for graphics applications, offering an automated solution with demonstrated performance gains.
The paper tackled the problem of GPU compiler optimization by developing a reinforcement learning framework to generate heuristics, resulting in improved frame rates for graphics applications, with an average uplift of 1.6% and up to 15.8% across benchmarks.
GPU compilers are complex software programs with many optimizations specific to target hardware. These optimizations are often controlled by heuristics hand-designed by compiler experts using time- and resource-intensive processes. In this paper, we developed a GPU compiler autotuning framework that uses off-policy deep reinforcement learning to generate heuristics that improve the frame rates of graphics applications. Furthermore, we demonstrate the resilience of these learned heuristics to frequent compiler updates by analyzing their stability across a year of code check-ins without retraining. We show that our machine learning-based compiler autotuning framework matches or surpasses the frame rates for 98% of graphics benchmarks with an average uplift of 1.6% up to 15.8%.