Search results for "GPU computing"
showing 8 items of 8 documents
Three-dimensional Fuzzy Kernel Regression framework for registration of medical volume data
2013
Abstract In this work a general framework for non-rigid 3D medical image registration is presented. It relies on two pattern recognition techniques: kernel regression and fuzzy c-means clustering. The paper provides theoretic explanation, details the framework, and illustrates its application to implement three registration algorithms for CT/MR volumes as well as single 2D slices. The first two algorithms are landmark-based approaches, while the third one is an area-based technique. The last approach is based on iterative hierarchical volume subdivision, and maximization of mutual information. Moreover, a high performance Nvidia CUDA based implementation of the algorithm is presented. The f…
3D grafikas atveidošana reālā laikā ar Vulkan API
2017
Maģistra darba mērķis veikt pētījumu par dažādām 3D grafikas atveidošanas iespējām uz mūsdienu datoriem, fokusējoties uz Vulkan programmsaskarni (turpmāk API), apskatot tās stiprās un vājas puses kā arī atšķirības no alternatīvām. Līdz šim tirgu ir dominējušas 2 API, lai izmantotu videokartes grafisko procesoru (turpmāk GPU) grafikas atveidošanai – OpenGL un DirectX. Tā kā mūsdienās 3D grafikas detalizācija un sarežģītība ir augoša, ir nepieciešami jauni risinājumi, kas precīzāk atbilst mūsdienu datoru arhitektūrai un ierīču daudzveidībai, kā arī dotu izstrādātājiem lielāku kontroli pār visiem notiekošajiem aprēķinu procesiem. Darba ietvaros ir izstrādāti divi 3D grafikas dzinēji. Vienā ir …
Future Processor Hardware Architectures for the Benefit of Precise Particle Accelerator Modeling
2017
Jaunās procesoru arhitektūras, kā grafiskie procesori (GPU) un Intel Many Integrated Cores (MIC) procesori, sniedz milzīgu veiktspējas potenciālu augstas veiktspējas skaitļošanas aplikācijās. Tomēr izstrādājot programmatūru, kas spēj izmantot šīs jaunās tehnoloģijas ir jāsaskarās ar dažādiem papildus grūtībām. Programmām ir jāspēj izmantot papildus paralēlisms, ko piedāvā šīs iekārtās, tām ir jāspēj pielāgoties dažādām procesoru arhitektūrām un jāizmanto dažādas izstrādes platformas, lai aplikācija spēdu darboties uz iekārtām no dažādiem ražotājiem. Dynamic Kernel Scheduler (DKS) tika izstrādāts, lai nodrošinātu papildus programmatūras slāni starp programmu un papildus processoriem. DKS nod…
Optical sectioning microscopy through single-shot Lightfield protocol
2020
Optical sectioning microscopy is usually performed by means of a scanning, multi-shot procedure in combination with non-uniform illumination. In this paper, we change the paradigm and report a method that is based in the light field concept, and that provides optical sectioning for 3D microscopy images after a single-shot capture. To do this we fi rst capture multiple orthographic perspectives of the sample by means of Fourier-domain integral microscopy (FiMic). The second stage of our protocol is the application of a novel refocusing algorithm that is able to produce optical sectioning in real time, and with no resolution worsening, in the case of sparse f luorescent samples.We provide the…
A GPU-accelerated augmented Lagrangian based L1-mean curvature Image denoising algorithm implementation
2015
This paper presents a graphics processing unit (GPU) implementation of a recently published augmented Lagrangian based L1-mean curvature image denoising algorithm. The algorithm uses a particular alternating direction method of multipliers to reduce the related saddle-point problem to an iterative sequence of four simpler minimization problems. Two of these subproblems do not contain the derivatives of the unknown variables and can therefore be solved point-wise without inter-process communication. Inparticular, this facilitates the efficient solution of the subproblem that deals with the non-convex term in the original objective function by modern GPUs. The two remaining subproblems are so…
On solving separable block tridiagonal linear systems using a GPU implementation of radix-4 PSCR method
2018
Partial solution variant of the cyclic reduction (PSCR) method is a direct solver that can be applied to certain types of separable block tridiagonal linear systems. Such linear systems arise, e.g., from the Poisson and the Helmholtz equations discretized with bilinear finite-elements. Furthermore, the separability of the linear system entails that the discretization domain has to be rectangular and the discretization mesh orthogonal. A generalized graphics processing unit (GPU) implementation of the PSCR method is presented. The numerical results indicate up to 24-fold speedups when compared to an equivalent CPU implementation that utilizes a single CPU core. Attained floating point perfor…
On GPU-accelerated fast direct solvers and their applications in image denoising
2015
GPU accelerated Monte Carlo simulations of lattice spin models
2011
We consider Monte Carlo simulations of classical spin models of statistical mechanics using the massively parallel architecture provided by graphics processing units (GPUs). We discuss simulations of models with discrete and continuous variables, and using an array of algorithms ranging from single-spin flip Metropolis updates over cluster algorithms to multicanonical and Wang-Landau techniques to judge the scope and limitations of GPU accelerated computation in this field. For most simulations discussed, we find significant speed-ups by two to three orders of magnitude as compared to single-threaded CPU implementations.