high performance computing on graphics processing units: hgpu.org

Modeling phenomena with networks has a wide application in many disciplines including biology, economics, sociology, and computer science. In network analysis modularity is an important measure for automatically extracting communities of closely connected nodes. Another important aspect of the network analysis is network visualization. Different techniques for network layout generation exist and the force-driven layout […]

Sparse matrix-vector multiplication (SpMV) is an important kernel in many scientific applications and is known to be memory bandwidth limited. On modern processors with wide SIMD and large numbers of cores, we identify and address several bottlenecks which may limit performance even before memory bandwidth: (a) low SIMD efficiency due to sparsity, (b) overhead due […]

We present results of numerical simulations of acoustic waves with the use of the Graphics Processing Unit (GPU) acceleration GAMER code which implements a second-order Godunov-type numerical scheme and adaptive mesh refinement (AMR). The AMR implementation is based on constructing a hierarchy of grid patches with an octree data structure. In this code a hybrid […]

Reliability is one of the most fundamental challenges for high performance computing (HPC) and cloud computing. Data replication is the de facto mechanism to achieve high reliability, even though it has been criticized for its high cost and low efficiency. Recent research showed promising results by switching the traditional data replication to a software-based RAID. […]

General-purpose computation on Graphics Processing Units (GPU) on personal computers has recently become an attractive alternative to parallel computing on clusters and supercomputers. We present the GPU-implementation of an accurate molecular dynamics algorithm for a system of spheres. The new hybrid CPU-GPU implementation takes into account all the degrees of freedom, including the quaternion representation […]

In this thesis work, we have mainly worked on two topics of GPU performance analysis. First, we have developed an analytical method and a timing estimation tool (TEG) to predict CUDA application’s performance for GT200 generation GPUs. TEG can predict GPU applications’ performance in cycle-approximate level. Second, we have developed an approach to estimate GPU […]

Given the tendency of creating interfaces between human and machines that increasingly allow simple ways of interaction, it is only natural that research effort is put into techniques that seek to simulate the most conventional mean of communication humans use: the speech. In the human auditory system, voice is automatically processed by the brain in […]

Developing accurate building energy simulation models to assist energy efficiency at speed and scale is one of the research goals of the Whole-Building and Community Integration group, which is a part of Building Technologies Research and Integration Center (BTRIC) at Oak Ridge National Laboratory (ORNL). The aim of the Autotune project is to speed up […]

This thesis reports on using dynamic load balancing methods on massively parallel computers in the context of multi-threaded computations. In particular we investigate the applicability of a randomized work stealing algorithm to ray tracing and breadth-first search as representatives of real-world applications with dynamic work creation. For our considerations we made use of current massively […]

HOOMD-blue is the first general purpose MD code built from the ground up for GPU acceleration, and has been actively developed since March 2007. It supports a variety of force fields and integrators targeted at soft-matter simulations. As an open source project, numerous developers have contributed useful feature additions back to the main code. High […]

As the demand increases for high performance and power efficiency in modern computer runtime systems and architectures, programmers are left with the daunting challenge of fully exploiting these systems for efficiency, high-level expressibility, and portability across different computing architectures. Emerging programming models such as the task-based runtime StarPU and many-core architectures such as GPUs force […]

Finance-related problems require more and more computations; therefore, the problem of finding efficient implementations for option pricing models on modern architectures has become an important challenge. Although there are numerous implementations of the Monte Carlo method on central processing units, many of them face limitations arising from the necessary increased computational power. In this paper, […]