Empirical performance modeling of GPU kernels using active learning

Prasanna Balaprakash, Karl Rupp, Azamat Mametjanov, Robert B. Gramacy, Paul D. Hovland, Stefan M. Wild
Mathematics and Computer Science Division, Argonne National Laboratory, Argonne, IL 60439
International Conference on Parallel Computing (ParCo), 2013


   title={Empirical Performance Modeling of GPU Kernels Using Active Learning},

   booktitle={International Conference on Parallel Computing – ParCo2013},


   address={Munich, Germany},

   author={Prasanna Balaprakash and K. Rupp and A. Mametjanov and R. B. Gramacy and Paul D. Hovland and S. M. Wild}


Download Download (PDF)   View View   Source Source   



We focus on a design-of-experiments methodology for developing empirical performance models of GPU kernels. Recently, we developed an iterative active learning algorithm that adaptively selects parameter configurations in batches for concurrent evaluation on CPU architectures in order to build performance models over the parameter space. In this paper, we illustrate the adoption of the algorithm when concurrent evaluations are not possible, which is particularly useful in the absence of GPU clusters. We present an empirical study of the algorithm on a diverse set of GPU kernels and hardware. We show that even when concurrent evaluations are not possible, the default batch mode of the algorithm yields better models and the iterative active learning algorithm reduces the overall time required to obtain high-quality empirical performance models for GPU kernels.
No votes yet.
Please wait...

* * *

* * *

HGPU group © 2010-2021 hgpu.org

All rights belong to the respective authors

Contact us: