Multitask and Transfer Learning for Autotuning Exascale Applications
This work addresses the challenge of efficiently tuning expensive exascale applications, offering a more suitable approach than existing methods, though it appears incremental as it builds on established multitask and transfer learning paradigms.
The paper tackled the problem of autotuning exascale applications by applying multitask and transfer learning to optimize performance parameters, resulting in an average 1.5x improvement in application runtime compared to state-of-the-art autotuners like OpenTuner and HpBandSter.
Multitask learning and transfer learning have proven to be useful in the field of machine learning when additional knowledge is available to help a prediction task. We aim at deriving methods following these paradigms for use in autotuning, where the goal is to find the optimal performance parameters of an application treated as a black-box function. We show comparative results with state-of-the-art autotuning techniques. For instance, we observe an average $1.5x$ improvement of the application runtime compared to the OpenTuner and HpBandSter autotuners. We explain how our approaches can be more suitable than some state-of-the-art autotuners for the tuning of any application in general and of expensive exascale applications in particular.