Corpus ID: 229156162

Learning to Schedule Halide Pipelines for the GPU

@article{Anderson2020LearningTS,
  title={Learning to Schedule Halide Pipelines for the GPU},
  author={L. Anderson and Andrew Adams and Karima Ma and Tzu-Mao Li and Jonathan Ragan-Kelley},
  journal={ArXiv},
  year={2020},
  volume={abs/2012.07145}
}
We present a new algorithm to automatically generate high-performance GPU implementations of complex imaging and machine learning pipelines, directly from high-level Halide algorithm code. It is fully automatic, requiring no schedule templates or hand-optimized kernels, and it targets a diverse range of computations which is significantly broader than existing autoschedulers. We address the scalability challenge of extending previous approaches to schedule large real world programs, while… Expand

Figures from this paper

References

SHOWING 1-10 OF 28 REFERENCES
Schedule Synthesis for Halide Pipelines on GPUs
  • 2
  • Highly Influential
Schedule Synthesis for Halide Pipelines through Reuse Analysis
  • 4
  • PDF
Automatically scheduling halide image processing pipelines
  • 88
  • PDF
TVM: An Automated End-to-End Optimizing Compiler for Deep Learning
  • 400
  • Highly Influential
  • PDF
Learning to optimize halide with tree search and random programs
  • 45
  • PDF
Differentiable programming for image processing and deep learning in halide
  • 36
  • PDF
PolyMage: Automatic Optimization for Image Processing Pipelines
  • 144
  • PDF
An effective fusion and tile size model for optimizing image processing pipelines
  • 21
...
1
2
3
...