Naoki Takada

Fast calculation of computer-generated-hologram on AMD HD5000 series GPU and OpenCL

In this paper, we report fast calculation of a computer-generated-hologram using a new architecture of the HD5000 series GPU (RV870) made by AMD and its new software development environment, OpenCL. Using a RV870 GPU and OpenCL, we can calculate 1,920 x 1,024 resolution of a CGH from a 3D object consisting of 1,024 points in 30 milli-seconds. The calculation speed realizes a speed approximately two times faster than that of a GPU made by NVIDIA.

Computational wave optics library for C++: CWO++ library

Abstract Diffraction calculations, such as the angular spectrum method and Fresnel diffractions, are used for calculating scalar light propagation. The calculations are used in wide-ranging optics fields: for example, Computer Generated Holograms (CGHs), digital holography, diffractive optical elements, microscopy, image encryption and decryption, three-dimensional analysis for optical devices and so on. However, increasing demands made by large-scale diffraction calculations have rendered the computational power of recent computers insufficient. We have already developed a numerical library for diffraction calculations using a Graphic Processing Unit (GPU), which was named the GWO library. However, this GWO library is not user-friendly, since it is based on C language and was also run only on a GPU. In this paper, we develop a new C++ class library for diffraction and CGH calculations, which is referred to as a CWO++ library, running on a CPU and GPU. We also describe the structure, performance, and usage examples of the CWO++ library. Program summary Program title: CWO++ Catalogue identifier: AELL_v1_0 Program summary URL: http://cpc.cs.qub.ac.uk/summaries/AELL_v1_0.html Program obtainable from: CPC Program Library, Queenʼs University, Belfast, N. Ireland Licensing provisions: Standard CPC licence, http://cpc.cs.qub.ac.uk/licence/licence.html No. of lines in distributed program, including test data, etc.: 109 809 No. of bytes in distributed program, including test data, etc.: 4 181 911 Distribution format: tar.gz Programming language: C++ Computer: General computers and general computers with NVIDIA GPUs Operating system: Windows XP, Vista, 7 Has the code been vectorized or parallelized?: Yes. 1 core processor used in CPU and many cores in GPU. RAM: 256 M bytes Classification: 18 External routines: CImg, FFTW Nature of problem: The CWO++ library provides diffraction calculations which are useful for Computer Generated Holograms (CGHs), digital holography, diffractive optical elements, microscopy, image encryption and decryption and three-dimensional analysis for optical devices. Solution method: FFT-based diffraction calculations, computer generated holograms by direct integration. Running time: The sample runs provided take approximately 5 minutes for the C++ version and 5 seconds for the C++ with GPUs version.

Fast high-resolution computer-generated hologram computation using multiple graphics processing unit cluster system.

To overcome the computational complexity of a computer-generated hologram (CGH), we implement an optimized CGH computation in our multi-graphics processing unit cluster system. Our system can calculate a CGH of 6,400×3,072 pixels from a three-dimensional (3D) object composed of 2,048 points in 55 ms. Furthermore, in the case of a 3D object composed of 4096 points, our system is 553 times faster than a conventional central processing unit (using eight threads).

Simplified electroholographic color reconstruction system using graphics processing unit and liquid crystal display projector.

We have constructed a simple color electroholography system that has excellent cost performance. It uses a graphics processing unit (GPU) and a liquid crystal display (LCD) projector. The structure of the GPU is suitable for calculating computer-generated holograms (CGHs). The calculation speed of the GPU is approximately 1,500 times faster than that of a central processing unit. The LCD projector is an inexpensive, high-performance device for displaying CGHs. It has high-definition LCD panels for red, green and blue. Thus, it can be easily used for color electroholography. For a three-dimensional object consisting of 1,000 points, our system succeeded in real-time color holographic reconstruction at rate of 30 frames per second.

Real-time color electroholography using multiple graphics processing units and multiple high-definition liquid-crystal display panels

We succeeded in developing a prototype three-dimensional (3-D) color television system by holography using three graphics processing units and three liquid-crystal display panels for displaying three primary colored images. A 3-D animation system featuring holography is said to have potential for ultimate 3-D television. Because of the technical complexity involved, however, the practical use of 3-D television is difficult. In our system, the image processing speed for holography is more than 300 times faster than that of today’s personal computers, and a 3-D image composed of 1000 points can be animated at virtually video rate, although the image size is as small as 5 cm3. The system succeeded in producing realistic 3-D color animation.

Numerical calculation library for diffraction integrals using the graphic processing unit: the GPU-based wave optics library

In optics, several diffraction integrals, such as the angular spectrum method and the Fresnel diffraction, are used for calculating scalar light propagation. The calculation result provides us with the optical characteristics of an optical device, the numerical reconstruction image from a hologram, and so forth. The acceleration of the calculation commonly uses the fast Fourier transform; however, in order to analyze a three-dimensional characteristic of an optical device and compute real-time reconstruction from holograms, recent computers do not have sufficient computational power. In this paper, we develop a numerical calculation library for the diffraction integrals using the graphic processing unit (GPU), the GWO library, and report the performance of the GWO library. The GPU chip allows us to use a highly parallel processor. The maximum computational speed of the GWO library is about 20 times faster than a personal computer.

Real-time spatiotemporal division multiplexing electroholography with a single graphics processing unit utilizing movie features

We propose a real-time spatiotemporal division multiplexing electroholography utilizing the features of movies. The proposed method spatially divides a 3-D object into plural parts and periodically selects a divided part in each frame, thereby reconstructing a three-dimensional (3-D) movie of the original object. Computer-generated holograms of the selected part are calculated by a single graphics processing unit and sequentially displayed on a spatial light modulator. Visual continuity enables a reconstructed movie of the original 3-D object. The proposed method realized a real-time reconstructed movie of a 3-D object composed of 11,646 points at over 30 frames per second (fps). We also displayed a reconstructed movie of a 3-D object composed of 44,647 points at about 10 fps.

High-speed FDTD simulation algorithm for GPU with compute unified device architecture

We proposed an FDTD algorithm for GPU with CUDA. Our GPU-FDTD algorithm performed high-speed FDTD simulation using GPU with CUDA, and maintained single-floating point accuracy. In the larger computational domain, the speedup factor becomes worse. The result suggests that the bottleneck of the FDTD simulation is memory bandwidth. Our GPU-FDTD algorithm can be applied to 3-D FDTD simulation. In future, we plan to implement our GPU-FDTD algorithm to the 3-D FDTD simulation.

Autoencoder-based holographic image restoration

We propose a holographic image restoration method using an autoencoder, which is an artificial neural network. Because holographic reconstructed images are often contaminated by direct light, conjugate light, and speckle noise, the discrimination of reconstructed images may be difficult. In this paper, we demonstrate the restoration of reconstructed images from holograms that record page data in holographic memory and quick response codes by using the proposed method.

Real-time electroholography using a multiple-graphics processing unit cluster system with a single spatial light modulator and the InfiniBand network

Abstract. Parallel calculations of large-pixel-count computer-generated holograms (CGHs) are suitable for multiple-graphics processing unit (multi-GPU) cluster systems. However, it is not easy for a multi-GPU cluster system to accomplish fast CGH calculations when CGH transfers between PCs are required. In these cases, the CGH transfer between the PCs becomes a bottleneck. Usually, this problem occurs only in multi-GPU cluster systems with a single spatial light modulator. To overcome this problem, we propose a simple method using the InfiniBand network. The computational speed of the proposed method using 13 GPUs (NVIDIA GeForce GTX TITAN X) was more than 3000 times faster than that of a CPU (Intel Core i7 4770) when the number of three-dimensional (3-D) object points exceeded 20,480. In practice, we achieved ∼40 tera floating point operations per second (TFLOPS) when the number of 3-D object points exceeded 40,960. Our proposed method was able to reconstruct a real-time movie of a 3-D object comprising 95,949 points.