Experimental Evaluation of Multiprecision Strategies for GMRES on GPUs

by   Jennifer A. Loe, et al.

Support for lower precision computation is becoming more common in accelerator hardware due to lower power usage, reduced data movement and increased computational performance. However, computational science and engineering (CSE) problems require double precision accuracy in several domains. This conflict between hardware trends and application needs has resulted in a need for multiprecision strategies at the linear algebra algorithms level if we want to exploit the hardware to its full potential while meeting the accuracy requirements. In this paper, we focus on preconditioned sparse iterative linear solvers, a key kernel in several CSE applications. We present a study of multiprecision strategies for accelerating this kernel on GPUs. We seek the best methods for incorporating multiple precisions into the GMRES linear solver; these include iterative refinement and parallelizable preconditioners. Our work presents strategies to determine when multiprecision GMRES will be effective and to choose parameters for a multiprecision iterative refinement solver to achieve better performance. We use an implementation that is based on the Trilinos library and employs Kokkos Kernels for performance portability of linear algebra kernels. Performance results demonstrate the promise of multiprecision approaches and demonstrate even further improvements are possible by optimizing low-level kernels.


page 1

page 6

page 7

page 9


A Study of Mixed Precision Strategies for GMRES on GPUs

Support for lower precision computation is becoming more common in accel...

Kokkos Kernels: Performance Portable Sparse/Dense Linear Algebra and Graph Kernels

As hardware architectures are evolving in the push towards exascale, dev...

Accelerating Geometric Multigrid Preconditioning with Half-Precision Arithmetic on GPUs

With the hardware support for half-precision arithmetic on NVIDIA V100 G...

Improving the Performance of the GMRES Method using Mixed-Precision Techniques

The GMRES method is used to solve sparse, non-symmetric systems of linea...

Least Squares on GPUs in Multiple Double Precision

This paper describes the application of the code generated by the CAMPAR...

Kernel methods through the roof: handling billions of points efficiently

Kernel methods provide an elegant and principled approach to nonparametr...

Performance Enhancement Strategies for Sparse Matrix-Vector Multiplication (SpMV) and Iterative Linear Solvers

Iterative solutions of sparse linear systems and sparse eigenvalue probl...

Please sign up or login with your details

Forgot password? Click here to reset