Abstract

Sparse tensors arise in problems in science, engineering, machine learning, and data analytics. Programs that operate on such tensors can exploit sparsity to reduce storage requirements and computational time. Developing and maintaining sparse software by hand, however, is a complex and error-prone task. Therefore, we propose treating sparsity as a property of tensors, not a tedious implementation task, and letting a sparse compiler generate sparse code automatically from a sparsity-agnostic definition of the computation. This paper discusses integrating this idea into MLIR.

Article

pdf

Movie

BibTeX

@article{bik2022,
  title={Compiler Support for Sparse Tensor Computations in MLIR},
  author={Aart J.C Bik and Penporn Koanantakool and Tatiana Shpeisman and Nicolas Vasilache and Bixia Zheng and Fredrik Kjolstad},
  journal={ACM Transactions on Architecture and Code Optimization},
  year={2022},
  month={September}
}