HyKKT: a hybrid direct-iterative method for solving KKT linear systems
- Stanford Univ., CA (United States)
- Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)
- Pacific Northwest National Laboratory (PNNL), Richland, WA (United States)
- Oak Ridge National Laboratory (ORNL), Oak Ridge, TN (United States)
Here, we propose a solution strategy for the large indefinite linear systems arising in interior methods for nonlinear optimization. The method is suitable for implementation on hardware accelerators such as graphical processing units (GPUs). The current gold standard for sparse indefinite systems is the LBLT factorization where L is a lower triangular matrix and B is 1×1 or 2×2 block diagonal. However, this requires pivoting, which substantially increases communication cost and degrades performance on GPUs. Our approach solves a large indefinite system by solving multiple smaller positive definite systems, using an iterative solver on the Schur complement and an inner direct solve (via Cholesky factorization) within each iteration. Cholesky is stable without pivoting, thereby reducing communication and allowing reuse of the symbolic factorization. We demonstrate the practicality of our approach on large optimal power flow problems and show that it can efficiently utilize GPUs and outperform LBLT factorization of the full system.
- Research Organization:
- Pacific Northwest National Laboratory (PNNL), Richland, WA (United States)
- Sponsoring Organization:
- USDOE National Nuclear Security Administration (NNSA); USDOE Office of Science (SC)
- Grant/Contract Number:
- AC05-76RL01830
- OSTI ID:
- 2446704
- Report Number(s):
- PNNL-SA--166808
- Journal Information:
- Optimization Methods and Software, Journal Name: Optimization Methods and Software Journal Issue: 2 Vol. 38; ISSN 1055-6788
- Publisher:
- Taylor & FrancisCopyright Statement
- Country of Publication:
- United States
- Language:
- English
Similar Records
Iterative algorithms for the normal and augemented equations
symPACK: A GPU-Capable Fan-Out Sparse Cholesky Solver