Search results for "Graphics processing unit"
showing 10 items of 42 documents
Nvidia CUDA parallel processing of large FDTD meshes in a desktop computer
2020
The Finite Difference in Time Domain numerical (FDTD) method is a well know and mature technique in computational electrodynamics. Usually FDTD is used in the analysis of electromagnetic structures, and antennas. However still there is a high computational burden, which is a limitation for use in combination with optimization algorithms. The parallelization of FDTD to calculate in GPU is possible using Matlab and CUDA tools. For instance, the simulation of a planar array, with a three dimensional FDTD mesh 790x276x588, for 6200 time steps, takes one day -elapsed time- using the CPU of an Intel Core i3 at 2.4GHz in a personal computer, 8Gb RAM. This time is reduced 120 times when the calcula…
Parallel Pairwise Epistasis Detection on Heterogeneous Computing Architectures
2016
This is a post-peer-review, pre-copyedit version of an article published in IEEE Transactions on Parallel and Distributed Systems. The final authenticated version is available online at: http://dx.doi.org/10.1109/TPDS.2015.2460247. [Abstract] Development of new methods to detect pairwise epistasis, such as SNP-SNP interactions, in Genome-Wide Association Studies is an important task in bioinformatics as they can help to explain genetic influences on diseases. As these studies are time consuming operations, some tools exploit the characteristics of different hardware accelerators (such as GPUs and Xeon Phi coprocessors) to reduce the runtime. Nevertheless, all these approaches are not able t…
GSaaS: A Service to Cloudify and Schedule GPUs
2018
Cloud technology is an attractive infrastructure solution that provides customers with an almost unlimited on-demand computational capacity using a pay-per-use approach, and allows data centers to increase their energy and economic savings by adopting a virtualized resource sharing model. However, resources such as graphics processing units (GPUs), have not been fully adapted to this model. Although, general-purpose computing on graphics processing units (GPGPU) is becoming more and more popular, cloud providers lack of flexibility to manage accelerators, because of the extended use of peripheral component interconnect (PCI) passthrough techniques to attach GPUs to virtual machines (VMs). F…
GPU-Based Optimisation of 3D Sensor Placement Considering Redundancy, Range and Field of View
2020
This paper presents a novel and efficient solution for the 3D sensor placement problem based on GPU programming and massive parallelisation. Compared to prior art using gradient-search and mixed-integer based approaches, the method presented in this paper returns optimal or good results in a fraction of the time compared to previous approaches. The presented method allows for redundancy, i.e. requiring selected sub-volumes to be covered by at least n sensors. The presented results are for 3D sensors which have a visible volume represented by cones, but the method can easily be extended to work with sensors having other range and field of view shapes, such as 2D cameras and lidars.
CUSHAW2-GPU: Empowering Faster Gapped Short-Read Alignment Using GPU Computing
2014
We present CUSHAW2-GPU to accelerate the CUSHAW2 algorithm using compute unified device architecture (CUDA)-enabled GPUs. Two critical GPU computing techniques, namely intertask hybrid CPU-GPU parallelism and tile-based Smith-Waterman map backtracking using CUDA, are investigated to facilitate fast alignments. By aligning both simulated and real reads to the human genome, our aligner yields comparable or better performance compared to BWA-SW, Bowtie2, and GEM. Furthermore, CUSHAW2-GPU with a Tesla K20c GPU achieves significant speedups over the multithreaded CUSHAW2, BWA-SW, Bowtie2, and GEM on the 12 cores of a high-end CPU for both single-end and paired-end alignment.
Iterative sparse matrix-vector multiplication for accelerating the block Wiedemann algorithm over GF(2) on multi-graphics processing unit systems
2012
SUMMARY The block Wiedemann (BW) algorithm is frequently used to solve sparse linear systems over GF(2). Iterative sparse matrix–vector multiplication is the most time-consuming operation. The necessity to accelerate this step is motivated by the application of BW to very large matrices used in the linear algebra step of the number field sieve (NFS) for integer factorization. In this paper, we derive an efficient CUDA implementation of this operation by using a newly designed hybrid sparse matrix format. This leads to speedups between 4 and 8 on a single graphics processing unit (GPU) for a number of tested NFS matrices compared with an optimized multicore implementation. We further present…
Parallelizing Epistasis Detection in GWAS on FPGA and GPU-Accelerated Computing Systems
2015
This is a post-peer-review, pre-copyedit version of an article published in IEEE - ACM Transactions on Computational Biology and Bioinformatics. The final authenticated version is available online at: http://dx.doi.org/10.1109/TCBB.2015.2389958 [Abstract] High-throughput genotyping technologies (such as SNP-arrays) allow the rapid collection of up to a few million genetic markers of an individual. Detecting epistasis (based on 2-SNP interactions) in Genome-Wide Association Studies is an important but time consuming operation since statistical computations have to be performed for each pair of measured markers. Computational methods to detect epistasis therefore suffer from prohibitively lon…
Towards an Efficient Implementation of an Accurate SPH Method
2020
A modified version of the Smoothed Particle Hydrodynamics (SPH) method is considered in order to overcome the loss of accuracy of the standard formulation. The summation of Gaussian kernel functions is employed, using the Improved Fast Gauss Transform (IFGT) to reduce the computational cost, while tuning the desired accuracy in the SPH method. This technique, coupled with an algorithmic design for exploiting the performance of Graphics Processing Units (GPUs), makes the method promising, as shown by numerical experiments.
Deep Learning-Based Methods for Prostate Segmentation in Magnetic Resonance Imaging
2021
Magnetic Resonance Imaging-based prostate segmentation is an essential task for adaptive radiotherapy and for radiomics studies whose purpose is to identify associations between imaging features and patient outcomes. Because manual delineation is a time-consuming task, we present three deep-learning (DL) approaches, namely UNet, efficient neural network (ENet), and efficient residual factorized convNet (ERFNet), whose aim is to tackle the fully-automated, real-time, and 3D delineation process of the prostate gland on T2-weighted MRI. While UNet is used in many biomedical image delineation applications, ENet and ERFNet are mainly applied in self-driving cars to compensate for limited hardwar…
Total Variation Regularization in Digital Breast Tomosynthesis
2013
We developed an iterative algebraic algorithm for the reconstruction of 3D volumes from limited-angle breast projection images. Algebraic reconstruction is accelerated using the graphics processing unit. We varied a total variation (TV)-norm parameter in order to verify the influence of TV regularization on the representation of small structures in the reconstructions. The Barzilai-Borwein algorithm is used to solve the inverse reconstruction problem. The quality of our reconstructions was evaluated with the Quart Mam/Digi Phantom, which features so-called Landolt ring structures to verify perceptibility limits. The evaluation of the reconstructions was done with an automatic LR detection a…