Combining multitask and transfer learning with deep Gaussian processes for autotuning-based performance engineering

doi:10.1177/10943420231166365

Combining multitask and transfer learning with deep Gaussian processes for autotuning-based performance engineering

2023-03-30

2

PRE

AI

P

Piotr Łuszczek *

W

Wissam M. Sid-Lakhdar

J

Jack Dongarra

DOI：10.1177/10943420231166365

Original

Original request for help

Save

Abstract

En 中文

We combine deep Gaussian processes (DGPs) with multitask and transfer learning for the performance modeling and optimization of HPC applications. Deep Gaussian processes merge the uncertainty quantification advantage of Gaussian processes (GPs) with the predictive power of deep learning. Multitask and transfer learning allow for improved learning efficiency when several similar tasks are to be learned simultaneously and when previous learned models are sought to help in the learning of new tasks, respectively. A comparison with state-of-the-art autotuners shows the advantage of our approach on two application problems. In this article, we combine DGPs with multitask and transfer learning to allow for both an improved tuning of an application parameters on problems of interest but also the prediction of parameters on any potential problem the application might encounter.

Keywords:

Gaussian process regression

performance autotuning

Efficient Global Optimization

Linear Coregionalization Model

Latin Hypercube Sampling

Combining multitask and transfer learning with deep Gaussian processes for autotuning-based performance engineering

Abstract

Journal

Organization