In the last three decades, a lot of research has been devoted to the optical response of an atomic media in near-to-resonant conditions and to how nonlinear optical properties are enhanced in these systems. However, a...
详细信息
In the last three decades, a lot of research has been devoted to the optical response of an atomic media in near-to-resonant conditions and to how nonlinear optical properties are enhanced in these systems. However, as current research turns its attention towards multi-level and multidimensional systems interacting with several electromagnetic fields, the ever-increasing complexity of these problems makes it difficult to treat the semiclassical model of the Maxwell-Bloch equations analytically without any strongly-limiting approximations. Thus, numerical methods and particularly robust and fast computational tools, capable of addressing such class of modern and future problems in photonics, are mandatory. In this paper, we describe the development and implementation of a Maxwell-Bloch numerical solver that exploits the massive parallelism of the GPUs to tackle efficiently problems in multidimensional settings or featuring Doppler broadening effects. This constitutes a simulation tool that is capable of addressing a vast class of problems with considerable reduction of simulation time, featuring speedups up to 15 compared with the same codes running on a CPU.
A software system has been developed for high-performance Computed Tomography (CT) reconstruction, simulation and other X-ray image processing tasks utilizing remote computer clusters optionally equipped with multiple...
详细信息
ISBN:
(纸本)9780819487513
A software system has been developed for high-performance Computed Tomography (CT) reconstruction, simulation and other X-ray image processing tasks utilizing remote computer clusters optionally equipped with multiple Graphics processingunits (GPUs). The system has a streamlined graphical User Interface for interaction with the cluster. Apart from extensive functionality related to X-ray CT in plane-wave and cone-beam forms, the software includes multiple functions for X-ray phase retrieval and simulation of phase-contrast imaging (propagation-based, analyzer crystal based and Talbot interferometry). Other features include several methods for image deconvolution, simulation of various phase-contrast microscopy modes (Zernike, Schlieren, Nomarski, dark-field, interferometry, etc.) and a large number of conventional image processing operations (such as FFT, algebraic and geometrical transformations, pixel value manipulations, simulated image noise, various filters, etc.). The architectural design of the system is described, as well as the two-level parallelization of the most computationally-intensive modules utilizing both the multiple CPU cores and multiple GPUs available in a local PC or a remote computer cluster. Finally, some results about the current system performance are presented. This system can potentially serve as a basis for a flexible toolbox for X-ray image analysis and simulation, that can efficiently utilize modern multi-processor hardware for advanced scientific computations.
Programmable graphics processingunits (GPUs) have emerged as excellent computational platforms for certain general-purpose applications. The data parallel execution capabilities of GPUs specifically point to the pote...
详细信息
ISBN:
(纸本)9781565553194
Programmable graphics processingunits (GPUs) have emerged as excellent computational platforms for certain general-purpose applications. The data parallel execution capabilities of GPUs specifically point to the potential for effective use in simulations of agent-based models (ABM). In this paper, the computational efficiency of ABM simulation on GPUs is evaluated on representative ABM benchmarks. The runtime speed of GPU-based models is compared to that of traditional CPU-based implementation, and also to that of equivalent models in traditional ABM toolkits (Repast and NetLogo). As expected, it is observed that, GPU-based ABM execution affords excellent speedup on simple models, with better speedup on models exhibiting good locality and fair amount of computation per memory element. Execution is two to three orders of magnitude faster with a GPU than with leading ABM toolkits, but at the cost of decrease in modularity, ease of programmability and reusability. At a more fundamental level, however, the data parallel paradigm is found to be somewhat at odds with traditional model-specification approaches for ABM. Effective use of data parallel execution, in general, seems to require resolution of modeling and execution challenges. Some of the challenges are identified and related solution approaches are described.
We outline an implementation of out-of-core direct solver for method of moments-based electromagnetic code. It is designed and executed on a computer cluster where each node has one or more graphical processing units....
详细信息
ISBN:
(纸本)9781479924318
We outline an implementation of out-of-core direct solver for method of moments-based electromagnetic code. It is designed and executed on a computer cluster where each node has one or more graphical processing units. The algorithm employs simultaneously multi-node parallelism of a cluster architecture and multi-processor parallelism of graphical processing units. This accelerates the overall analysis and allows solving larger problems. The complexity of the solvable problems is limited only with computer resources of the cluster. We perform numerical experiments on a cluster with 8 nodes and with 2 graphical processing units per node. The achieved efficiency of multi-node solution is around 80%. As an illustration of the efficiency of the approach, the radiation pattern of a microstirp antenna on a helicopter fuselage, and monostatic radar cross section of a global hawk are provided, as well as the simulation times.
暂无评论