Evaluation of the deflated preconditioned CG method to solve bubbly and porous media flow problems on GPU and CPU

Gupta, Rohit; Lukarski, Dimitar; Gijzen, Martin B. van; Vuik, C.

doi:10.1002/fld.4170

Cited by 2 publications

(1 citation statement)

References 14 publications

(24 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Currently, GPUs are popular for many engineering applications [23,24]. With a high computational load in the fully distributed parts of the method due to full local reconstruction and projection matrices, and with low interpartition communication for the PME procedures, the D-SABRE method was conceptualized for implementation on parallel hardware, specifically on GPUs.…”

Section: Adaptation Of D-sabre For Gpusmentioning

confidence: 99%

GPU implementation for spline-based wavefront reconstruction

et al. 2018

View full text Add to dashboard Cite

This paper presents an adaptation of the distributed-spline-based aberration reconstruction method for Shack-Hartmann (SH) slope measurements to extremely large-scale adaptive optics systems and the execution on graphics processing units (GPUs). The introduction of a hierarchical multi-level scheme for the elimination of piston offsets between the locally computed wavefront (WF) estimates solves the piston error propagation observed for a large number of partitions with the original version. To obtain a fully distributed method for WF correction, the projection of the phase estimates is locally approximated and applied in a distributed fashion, providing stable results for low and medium actuator coupling. An implementation of the method with the parallel computing platform CUDA exploits the inherently distributed nature of the algorithm. With a standard off-the-shelf GPU, the computation of the adaptive optics correction updates is accomplished in under 1 ms for the benchmark case of a 200×200 SH array.

show abstract

Section: Adaptation Of D-sabre For Gpusmentioning

confidence: 99%

GPU implementation for spline-based wavefront reconstruction

et al. 2018

View full text Add to dashboard Cite

show abstract

Deflated Preconditioned Conjugate Gradients for Nonlinear Diffusion Image Enhancement

Shan

Gijzen

2020

Lecture Notes in Computational Science and Engineering

View full text Add to dashboard Cite

Nonlinear diffusion equations have been successfully used for image enhancement by reducing the noise in the image while protecting the edges. In discretized form, the denoising requires the solution of a sequence of linear systems. The underlying system matrices stem from a discrete diffusion operator with large jumps in the diffusion coefficients. As a result these matrices can be very illconditioned, which leads to slow convergence for iterative methods such as the Conjugate Gradient method. To speed-up the convergence we use deflation and preconditioning. The deflation vectors are defined by a decomposition of the image. The resulting numerical method is easy to implement and matrix-free. We evaluate the performance of the method on a simulated image and on a measured low-field MR image for various types of deflation vectors.

show abstract

Evaluation of the deflated preconditioned CG method to solve bubbly and porous media flow problems on GPU and CPU

Cited by 2 publications

References 14 publications

GPU implementation for spline-based wavefront reconstruction

GPU implementation for spline-based wavefront reconstruction

Deflated Preconditioned Conjugate Gradients for Nonlinear Diffusion Image Enhancement

Contact Info

Product

Resources

About