In this work we study the implementations of deflation and preconditioning techniques for solving ill-conditioned linear systems using iterative methods. Solving such systems can be a time-consuming process because of the jumps in the coefficients due to large difference in material properties. We have developed implementations of the iterative methods with these preconditioning techniques on […]

October 11, 2015 by hgpu

Hardware-aware design and optimization is crucial in exploiting emerging architectures for PDE-based computational fluid dynamics applications. In this work, we study optimizations aimed at acceleration of OpenFOAM-based applications on emerging hybrid heterogeneous platforms. OpenFOAM uses MPI to provide parallel multi-processor functionality, which scales well on homogeneous systems but does not fully utilize the potential per-node […]

May 29, 2015 by hgpu

Lattice Quantum Chromodynamics simulations typically spend most of the runtime in inversions of the Fermion Matrix. This part is therefore frequently optimized for various HPC architectures. Here we compare the performance of the Intel Xeon Phi to current Kepler-based NVIDIA Tesla GPUs running a conjugate gradient solver. By exposing more parallelism to the accelerator through […]

November 18, 2014 by hgpu

We revisit the implementation of iterative solvers on discrete graphics processing units and demonstrate the benefit of implementations using extensive kernel fusion for pipelined formulations over conventional implementations of classical formulations. The proposed implementations with both CUDA and OpenCL are freely available in ViennaCL and achieve up to three-fold performance gains when compared to other […]

October 16, 2014 by hgpu

The runtime of a Lattice QCD simulation is dominated by a small kernel, which calculates the product of a vector by a sparse matrix known as the "Dslash" operator. Therefore, this kernel is frequently optimized for various HPC architectures. In this contribution we compare the performance of the Intel Xeon Phi to current Kepler-based NVIDIA […]

September 5, 2014 by hgpu

The main purpose of this work is to show the clear advantages of using modern parallel tools in solving the SLAE. The BiCGStab method was used for solving system of the linear equations. This paper contains some details about this method. To accelerate computations on the GPU several technologies (such as CUBLAS, OpenACC) were used. […]

June 14, 2014 by hgpu

The main purpose of this work is to show the advantages of using various approaches of heterogeneous programming. The results were received on the example of solving the system of the linear equations by the conjugate gradient method. High-level and low-level technologies (OpenACC and CUDA respectively) were used to accelerate computations on the GPU. The […]

May 29, 2014 by hgpu

The paper presents a highly efficient way of simulating the dynamic behavior of deformable objects by means of the finite element method (FEM) with computations performed on Graphics Processing Units (GPU). The presented implementation reduces bottlenecks related to memory accesses by grouping the necessary data per node pairs, in contrast to the classical way done […]

April 14, 2014 by hgpu

We present a performance analysis of a parallel implementation of both conjugate gradient and preconditioned conjugate gradient solvers using graphic processing units with CUDA parallel programming model. The solvers were optimized for a fast solution of sparse systems of equations arising from Finite Element Analysis (FEA) of electromagnetic phenomena. The preconditioners were Incomplete Cholesky factorization […]

March 12, 2014 by hgpu

The progress of high performance computing platforms is dramatic, and most of the simulations carried out on these platforms, result in improvements on one level, yet exposes shortcomings of the current CFD packages capabilities. Therefore, hardware-aware design and optimizations are crucial towards exploiting the modern computing resources. This thesis proposes optimizations aimed at acceleration numerical […]

December 21, 2013 by hgpu

Matrix solvers play a crucial role in solving real world physics problem. In engineering practice, transition analysis is most often used, which requires a series of similar matrices to be solved. However, any specific solver with/without preconditioner cannot achieve high performance gain for all matrices. This paper recommends Conjugate Gradient iterative solver with SSOR approximate […]

September 20, 2013 by hgpu

Graphics Processing Unit (GPU) has obtained great success in scientific computations for its tremendous computational horsepower and very high memory bandwidth. This paper discusses the efficient way to implement polynomial preconditioned conjugate gradient solver for the finite element computation of elasticity on NVIDIA GPUs using Compute Unified Device Architecture (CUDA). Sliced Block ELLPACK (SBELL) format […]

September 14, 2013 by hgpu