high performance computing on graphics processing units: hgpu.org

The numerical solution of two-layer shallow water systems is required to simulate accurately stratified fluids, which are ubiquitous in nature: they appear in atmospheric flows, ocean currents, oil spills, etc. Moreover, the implementation of the numerical schemes to solve these models in realistic scenarios imposes huge demands of computing power. In this paper, we tackle […]

We introduce a parallel implementation of the Simple Linear Iterative Clustering (SLIC) superpixel segmentation. Our implementation uses GPU and the NVIDIA CUDA framework. Using a single graphic card, our implementation achieves speedups of 10x~20x from the sequential implementation. This allow us to use the superpixel segmentation method in real-time performance. Our implementation is compatible with […]

Recent GPU developments have attracted much interest in the HPC community. Since each GPU interface requires a dedicated host processor, the unused high performance non-GPU processors are simply wasted. GPUs are energy intensive and are more likely to fail than CPUs, we are interested in using all processors to a) boosting application performance, and b) […]

This paper presents application of several graph algorithms for comparison and visualization of real-world networks. In order to obtain interactive and robust framework for analysis of large graphs we use CUDA implementations of all-shortest-paths (APSP) and breadth-first-search (BFS) algorithms along with CULA matrix decomposition routines. Such an approach allows for efficient computation of graph feature […]

Multicore computational accelerators such as GPUs are now commodity components for high-performance computing at scale. While such accelerators have been studied in some detail as stand-alone computational engines, their integration in large-scale distributed systems raises new challenges and trade-offs. In this paper, we present an exploration of resource management alternatives for building asymmetric accelerator-based distributed […]

In this work, we present our implementation of the density functional theory (DFT) plane wave pseudopotential (PWP) calculations on GPU clusters. This GPU version is developed based on a CPU DFT-PWP code: PEtot, which can calculate up to a thousand atoms on thousands of processors. Our test indicates that the GPU version can have a […]

The proposed paper will present details on recent advancements in scientific computing in terms of integrating new hardware and software to greatly enhance the computational efficiency of comprehensive rotorcraft analysis. The focus will be on showing the tremendous computational accelerations that are possible (i.e., orders of magnitude speed up) by using software developments in the […]

We present a fully multi-GPU-based double-precision solver for the three-dimensional two-phase incompressible Navier-Stokes equations. An in-depth performance analysis shows a realistic speed-up of the order of three by comparing equally priced GPUs and CPUs and more than a doubling in energy efficiency for GPUs. We observe profound strong and weak scaling on a multi-GPU cluster.

We perform hybrid Monte Carlo (HMC) simulatons of two flavors QCD with the optimal domain-wall fermion (ODWF) on the $ 16^3 times 32 $ lattice (with lattice spacing $ a sim 0.1 $ fm), for eight sea-quark masses corresponding to pion masses in the range 230-580 MeV. We calculate the mass and the decay constant […]

The Simplex algorithm is a well known method to solve linear programming (LP) problems. In this paper, we propose an implementation via CUDA of the Simplex method on a multi GPU architecture. Computational tests have been carried out on randomly generated instances for non-sparse LP problems. The tests show a maximum speedup of 24:5 with […]

We deal with the development of a computationally effective approach for the synthesis of equivalently tapered, aperiodic linear arrays, i.e. arrays matching the requirements on the power pattern by acting only on the element positions and excitation phases. The computational effectiveness of the algorithm is reached by the development of a parallel Non Uniform Fast […]

This paper presents an All-in-One video analytics system, a compact, multi-channel, real-time, video monitoring, event detection, alarm notification, event recording and browsing solution implemented on low cost hardware, taking advantage of NVIDIA’s GPU CUDA platform. An inventive distribution of video object detection and tracking processing chain between the GPUs and the CPU provides maximum efficiency […]