A Randomized Algorithm for Preconditioner Selection

08/01/2019
by   Conner DiPaolo, et al.
0

The task of choosing a preconditioner M to use when solving a linear system Ax=b with iterative methods is difficult. For instance, even if one has access to a collection M_1,M_2,...,M_n of candidate preconditioners, it is currently unclear how to practically choose the M_i which minimizes the number of iterations of an iterative algorithm to achieve a suitable approximation to x. This paper makes progress on this sub-problem by showing that the preconditioner stability I-M^-1A_F, known to forecast preconditioner quality, can be computed in the time it takes to run a constant number of iterations of conjugate gradients through use of sketching methods. This is in spite of folklore which suggests the quantity is impractical to compute, and a proof we give that ensures the quantity could not possibly be approximated in a useful amount of time by a deterministic algorithm. Using our estimator, we provide a method which can provably select the minimal stability preconditioner among n candidates using floating point operations commensurate with running on the order of n n steps of the conjugate gradients algorithm. Our method can also advise the practitioner to use no preconditioner at all if none of the candidates appears useful. The algorithm is extremely easy to implement and trivially parallelizable. In one of our experiments, we use our preconditioner selection algorithm to create to the best of our knowledge the first preconditioned method for kernel regression reported to never use more iterations than the non-preconditioned analog in standard tests.

READ FULL TEXT
research
10/01/2019

ARCHITECT: Arbitrary-precision Hardware with Digit Elision for Efficient Iterative Compute

Many algorithms feature an iterative loop that converges to the result o...
research
05/17/2021

A deterministic Kaczmarz algorithm for solving linear systems

We propose a deterministic Kaczmarz method for solving linear systems A=...
research
03/04/2020

Stability Analysis of Inline ZFP Compression for Floating-Point Data in Iterative Methods

Currently, the dominating constraint in many high performance computing ...
research
04/16/2019

Short Overview of Early Developments of the Hardy Cross Type Methods for Computation of Flow Distribution in Pipe Networks

Hardy Cross originally proposed a method for analysis of flow in network...
research
11/13/2020

An augmented wavelet reconstructor for atmospheric tomography

Atmospheric tomography, i.e. the reconstruction of the turbulence profil...
research
05/07/2019

Hiring Under Uncertainty

In this paper we introduce the hiring under uncertainty problem to model...
research
05/06/2022

Conditions for Digit Stability in Iterative Methods Using the Redundant Number Representation

Iterative methods play an important role in science and engineering appl...

Please sign up or login with your details

Forgot password? Click here to reset