A GPU-Accelerated Linear Solver for Massively Parallel Underground Simulations

Modern engineering applications require the solution of linear systems of millions or even billions of equations. The solution of the linear system takes most of the simulation for large scale simulations, and represent the bottleneck in developing scientific and technical software. Usually, preconditioned iterative solvers are preferred because of their low memory requirements and they can have a high level of parallelism. Approximate inverses have been proven to be robust and effective preconditioners in several contexts. In this communication, we present an adaptive Factorized Sparse Approximate Inverse (FSAI) preconditioner with a very high level of parallelism in both set-up and application. Its inherent parallelism makes FSAI an ideal candidate for a GPU-accelerated implementation, even if taking advantage of this hardware is not a trivial task, especially in the set-up stage. An extensive numerical experimentation has been performed on industrial underground applications. It is shown that the proposed approach outperforms more traditional preconditioners in challenging underground simulation, greatly reducing time-to-solution.

A Comparative Study of Block Incomplete Sparse Approximate Inverses Preconditioning on Tesla K20 and V100 GPUs

Incomplete Sparse Approximate Inverses (ISAI) has shown some advantages over sparse triangular solves on GPUs when it is used for the incomplete LU based preconditioner. In this paper, we extend the single GPU method for Block–ISAI to multiple GPUs algorithm by coupling Block–Jacobi preconditioner, and introduce the detailed implementation in the open source numerical package PETSc. In the experiments, two representative cases are performed and a comparative study of Block–ISAI on up to four GPUs are conducted on two major generations of NVIDIA’s GPUs (Tesla K20 and Tesla V100). Block–Jacobi preconditioning with Block–ISAI (BJPB-ISAI) shows an advantage over the level-scheduling based triangular solves from the cuSPARSE library for the cases, and the overhead of setting up Block–ISAI and the total wall clock times of GMRES is greatly reduced using Tesla V100 GPUs compared to Tesla K20 GPUs.

GPU Preconditioning for Block Linear Systems Using Block Incomplete Sparse Approximate Inverses

Solving sparse triangular systems is the building block for incomplete LU- (ILU-) based preconditioning, but parallel algorithms, such as the level-scheduling scheme, are sometimes limited by available parallelism extracted from the sparsity pattern. In this study, the block version of the incomplete sparse approximate inverses (ISAI) algorithm is studied, and the block-ISAI is considered for preconditioning by proposing an efficient algorithm and implementation on graphical processing unit (GPU) accelerators. Performance comparisons are carried out between the proposed algorithm and serial and parallel block triangular solvers from PETSc and cuSPARSE libraries. The experimental results show that GMRES (30) with the proposed block-ISAI preconditioning achieves accelerations 1.4 × –6.9 × speedups over that using the cuSPARSE library on NVIDIA Tesla V100 GPU.

A GPU-accelerated adaptive FSAI preconditioner for massively parallel simulations

The solution of linear systems of equations is a central task in a number of scientific and engineering applications. In many cases the solution of linear systems may take most of the simulation time thus representing a major bottleneck in the further development of scientific and technical software. For large scale simulations, nowadays accounting for several millions or even billions of unknowns, it is quite common to resort to preconditioned iterative solvers for exploiting their low memory requirements and, at least potential, parallelism. Approximate inverses have been shown to be robust and effective preconditioners in various contexts. In this work, we show how adaptive Factored Sparse Approximate Inverse (aFSAI), characterized by a very high degree of parallelism, can be successfully implemented on a distributed memory computer equipped with GPU accelerators. Taking advantage of GPUs in adaptive FSAI set-up is not a trivial task, nevertheless we show through an extensive numerical experimentation how the proposed approach outperforms more traditional preconditioners and results in a close-to-ideal behavior in challenging linear algebra problems.

Nonfragile H ∞ Stabilizing Nonlinear Systems Described by Multivariable Hammerstein Models

This paper presents the problem of robust and nonfragile stabilization of nonlinear systems described by multivariable Hammerstein models. The objective is focused on the design of a nonfragile feedback controller such that the resulting closed-loop system is globally asymptotically stable with robust H ∞ disturbance attenuation in spite of controller gain variations. First, the parameters of linear and nonlinear blocks characterizing the multivariable Hammerstein model structure are separately estimated by using a subspace identification algorithm. Second, approximate inverse nonlinear functions of polynomial form are proposed to deal with nonbijective invertible nonlinearities. Thereafter, the Takagi–Sugeno model representation is used to decompose the composition of the static nonlinearities and their approximate inverses in series with the linear subspace dynamic submodel into linear fuzzy parts. Besides, sufficient stability conditions for the robust and nonfragile controller synthesis based on quadratic Lyapunov function, H ∞ criterion, and linear matrix inequality approach are provided. Finally, a numerical example based on twin rotor multi-input multi-output system is considered to demonstrate the effectiveness.

Low-rank Updates of Preconditioners for Sequences of Symmetric Linear Systems

The aim of this survey is to review some recent developements in devising efficient preconditioners for sequences of linear systems A x = b. Such a problem arise in many scientific applications, such as discretization of transient PDEs, solution of eigenvalue problems, (Inexact) Newton method applied to nonlinear systems, rational Krylov methods for computing a function of a matrix. Full purpose preconditioners such as the Incomplete Cholesky (IC) factorization or approximate inverses are aimed at clustering eigenvalues of the preconditioned matrices around one. In this paper we will analyze a number of techniques of updating a given IC preconditioner (which we denote as P0 in the sequel) by a low-rank matrix with the aim of further improving this clustering. The most popular low-rank strategies are aimed at removing the smallest eigenvalues (deflation) or at shifting them towards the middle of the spectrum. The low-rank correction is based on a (small) number of linearly independent vectors whose choice is crucial for the effectiveness of the approach. In many cases these vectors are approximations of eigenvectors corresponding to the smallest eigenvalues of the preconditioned matrix P0 A. We will also review some techniques to efficiently approximate these vectors when incorporated within a sequence of linear systems all possibly having constant (or slightly changing) coefficient matrices. Numerical results concerning sequences arising from discretization of linear/nonlinear PDEs and iterative solution of eigenvalue problems show that the performance of a given iterative solver can be very much enhanced by the use of low-rank updates.

Efficient Algebraic Multigrid Preconditioners on Clusters of GPUs

Many scientific applications require the solution of large and sparse linear systems of equations using Krylov subspace methods; in this case, the choice of an effective preconditioner may be crucial for the convergence of the Krylov solver. Algebraic MultiGrid (AMG) methods are widely used as preconditioners, because of their optimal computational cost and their algorithmic scalability. The wide availability of GPUs, now found in many of the fastest supercomputers, poses the problem of implementing efficiently these methods on high-throughput processors. In this work we focus on the application phase of AMG preconditioners, and in particular on the choice and implementation of smoothers and coarsest-level solvers capable of exploiting the computational power of clusters of GPUs. We consider block-Jacobi smoothers using sparse approximate inverses in the solve phase associated with the local blocks. The choice of approximate inverses instead of sparse matrix factorizations is driven by the large amount of parallelism exposed by the matrix-vector product as compared to the solution of large triangular systems on GPUs. The selected smoothers and solvers are implemented within the AMG preconditioning framework provided by the MLD2P4 library, using suitable sparse matrix data structures from the PSBLAS library. Their behaviour is illustrated in terms of execution speed and scalability, on a test case concerning groundwater modelling, provided by the Jülich Supercomputing Center within the Horizon 2020 Project EoCoE.

A Class of Symmetric Factored Approximate Inverses and Hybrid Two-Level Solver

A new class of symmetric factored approximate inverses is proposed and used in conjunction with the Preconditioned Conjugate Gradient method for solving sparse symmetric linear systems. Additionally, a new hybrid two-level solver is proposed utilizing a block independent set reordering, in order to create the two level hierarchy. The Schur complement is formed explicitly by inverting the blocks created by reordering. Then, the preconditioned conjugate gradient method is used in conjunction with the symmetric factored approximate inverse to solve the reduced order linear system. Furthermore, numerical results on the performance and convergence behavior for solving various model problems are presented.