A Two-Tier Design Space Exploration Algorithm to Construct a GPU Performance Predictor

Abstract

Graphics Processing Units (GPUs) have a large and complex design space that needs to be explored in order to optimize the performance of future GPUs. Statistical techniques are useful tools to help computer architects to predict performance of complex processors. In this study, these methods are utilized to build an effective performance prediction model for a Fermi GPU. The design space of this GPU is more than 8 million points. In order to build an accurate model, we propose a two-tier algorithm which builds a multiple linear regression model from a small set of simulated data. In this algorithm the Plackett and Burman design is used to find the key parameters of the GPU, and further simulations are guided by a fractional factorial design for the most important parameters. The generated performance model is able to predict the performance of any other point in the design space with an average prediction error between 1% to 5% for different benchmark applications. This accuracy is achieved by only sampling between 0.0003% to 0.0015% of the full design space.