Saved in:
Bibliographic Details
Main Authors: Regev, Shaked, Chiang, Nai-Yuan, Darve, Eric, Petra, Cosmin G., Saunders, Michael A., Świrydowicz, Kasia, Peleš, Slaven
Format: Preprint
Published: 2021
Subjects:
Online Access:https://arxiv.org/abs/2110.03636
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912945818042368
author Regev, Shaked
Chiang, Nai-Yuan
Darve, Eric
Petra, Cosmin G.
Saunders, Michael A.
Świrydowicz, Kasia
Peleš, Slaven
author_facet Regev, Shaked
Chiang, Nai-Yuan
Darve, Eric
Petra, Cosmin G.
Saunders, Michael A.
Świrydowicz, Kasia
Peleš, Slaven
contents We propose a solution strategy for linear systems arising in interior method optimization, which is suitable for implementation on hardware accelerators such as graphical processing units (GPUs). The current gold standard for solving these systems is the LDL^T factorization. However, LDL^T requires pivoting during factorization, which substantially increases communication cost and degrades performance on GPUs. Our novel approach solves a large indefinite system by solving multiple smaller positive definite systems, using an iterative solve for 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 and show that on large systems it can efficiently utilize GPUs and outperform LDL^T factorization of the full system.
format Preprint
id arxiv_https___arxiv_org_abs_2110_03636
institution arXiv
publishDate 2021
record_format arxiv
spellingShingle A Hybrid Direct-Iterative Method for Solving KKT Linear Systems
Regev, Shaked
Chiang, Nai-Yuan
Darve, Eric
Petra, Cosmin G.
Saunders, Michael A.
Świrydowicz, Kasia
Peleš, Slaven
Optimization and Control
Distributed, Parallel, and Cluster Computing
15, 65, 68
G.1
We propose a solution strategy for linear systems arising in interior method optimization, which is suitable for implementation on hardware accelerators such as graphical processing units (GPUs). The current gold standard for solving these systems is the LDL^T factorization. However, LDL^T requires pivoting during factorization, which substantially increases communication cost and degrades performance on GPUs. Our novel approach solves a large indefinite system by solving multiple smaller positive definite systems, using an iterative solve for 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 and show that on large systems it can efficiently utilize GPUs and outperform LDL^T factorization of the full system.
title A Hybrid Direct-Iterative Method for Solving KKT Linear Systems
topic Optimization and Control
Distributed, Parallel, and Cluster Computing
15, 65, 68
G.1
url https://arxiv.org/abs/2110.03636