Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

GPU Accelerated Sparse Cholesky Factorization

Conference ·

The solution of sparse symmetric positive definite linear systems is an important computational kernel in large-scale scientific and engineering modeling and simulation. We will solve the linear systems using a direct method, in which a Cholesky factorization of the coefficient matrix is performed using a right-looking approach and the resulting triangular factors are used to compute the solution. Sparse Cholesky factorization is compute intensive. In this work we investigate techniques for reducing the factorization time in sparse Cholesky factorization by offloading some of the dense matrix operations on a GPU. We will describe the techniques we have considered. We achieved up to 4x speedup compared to the CPU-only version.

Research Organization:
Lawrence Berkeley National Laboratory (LBNL), Berkeley, CA (United States)
Sponsoring Organization:
US Department of Energy; USDOE Office of Science (SC)
DOE Contract Number:
AC02-05CH11231
OSTI ID:
2586562
Resource Relation:
Proceedings of Sc 2024 W Workshops of the International Conference for High Performance Computing Networking Storage and Analysis
Country of Publication:
United States
Language:
English

Similar Records

symPACK: A GPU-Capable Fan-Out Sparse Cholesky Solver
Conference · Sun Nov 12 23:00:00 EST 2023 · OSTI ID:2006802

Sparse Cholesky factorization on a multiprocessor
Thesis/Dissertation · Wed Dec 31 23:00:00 EST 1986 · OSTI ID:7183866

A scalable parallel algorithm for sparse Cholesky factorization
Book · Fri Dec 30 23:00:00 EST 1994 · OSTI ID:87680

Related Subjects