2016 International Conference on Parallel Architecture and Compilation Techniques (PACT) (2016)

Haifa, Israel

Sept. 11, 2016 to Sept. 15, 2016

ISBN: 978-1-5090-5308-7

pp: 247-259

Hongbo Rong , Parallel Computing Lab, Intel Corporation, United States

Jongsoo Park , Parallel Computing Lab, Intel Corporation, United States

Lingxiang Xiang , Parallel Computing Lab, Intel Corporation, United States

Todd A. Anderson , Parallel Computing Lab, Intel Corporation, United States

Mikhail Smelyanskiy , Parallel Computing Lab, Intel Corporation, United States

ABSTRACT

The sparse matrix is a key data structure in various domains such as high-performance computing, machine learning, and graph analytics. To maximize performance of sparse matrix operations, it is especially important to optimize across the operations and not just within individual operations. While a straightforward per-operation mapping to library routines misses optimization opportunities, manually optimizing across the boundary of library routines is time-consuming and error-prone, sacrificing productivity.

INDEX TERMS

Sparse matrices, Optimization, Context, Libraries, Linear algebra, Symmetric matrices, Productivity

CITATION

H. Rong, J. Park, L. Xiang, T. A. Anderson and M. Smelyanskiy, "Sparso: Context-driven optimizations of sparse linear algebra,"

*2016 International Conference on Parallel Architecture and Compilation Techniques (PACT)*, Haifa, Israel, 2016, pp. 247-259.

doi:10.1145/2967938.2967943

CITATIONS

SEARCH