OpenTuner
Jason Ansel, Shoaib Kamil, Kalyan K. Veeramachaneni et al. · International Conference on Parallel Architectures and Compilation Techniques · 2014
Program autotuning has been shown to achieve better or more portable performance in a number of domains. However, autotuners themselves are rarely portable between projects, for a number of reasons: using a domain-informed search space representation is critical to achieving good results; search spaces can be intractably large and require advanced machine learning techniques; and the landscape of search spaces can vary greatly between different problems, sometimes requiring domain specific search techniques to explore efficiently.
Cited by 530 works. doi:10.1145/2628071.2628092 Open-access PDF
Authors
- Jason Ansel
- Shoaib Kamil
- Kalyan K. Veeramachaneni
- Jonathan Ragan‐Kelley
- Jeffrey Bosboom
- Una-May O’Reilly
- Saman P. Amarasinghe
References
- The Design and Implementation of FFTW3
- The SPLASH-2 programs
- Automatically Tuned Linear Algebra Software
- Top500 Supercomputer Sites
- The LINPACK Benchmark: past, present and future
- Halide
- OSKI: A library of automatically tuned sparse matrix kernels
- Green
- Using Machine Learning to Focus Iterative Optimization
- Decoupling algorithms from schedules for easy optimization of image processing pipelines