Details

Autor(en) / Beteiligte

• Piedrahita-Quintero, Pablo
• Trujillo, Carlos
• Garcia-Sucerquia, Jorge

Titel

JDiffraction: A GPGPU-accelerated JAVA library for numerical propagation of scalar wave fields

Ist Teil von

• Computer physics communications, 2017-05, Vol.214, p.128-139

Ort / Verlag

Elsevier B.V

Erscheinungsjahr

2017

Quelle

Alma/SFX Local Collection

Beschreibungen/Notizen

• JDiffraction, a GPGPU-accelerated JAVA library for numerical propagation of scalar wave fields, is presented. Angular spectrum, Fresnel transform, and Fresnel–Bluestein transform are the numerical algorithms implemented in the methods and functions of the library to compute the scalar propagation of the complex wavefield. The functionality of the library is tested with the modeling of easy to forecast numerical experiments and also with the numerical reconstruction of a digitally recorded hologram. The performance of JDiffraction is contrasted with a library written for C++, showing great competitiveness in the apparently less complex environment of JAVA language. JDiffraction also includes JAVA easy-to-use methods and functions that take advantage of the computation power of the graphic processing units to accelerate the processing times of 2048×2048 pixel images up to 74 frames per second. Program title: JDiffraction Program Files doi:http://dx.doi.org/10.17632/nwrwz7mn7h.1 Licensing provisions: GNU General Public License 3 (GPL) Programming language: JAVA Nature of problem: In order to perform the numerical propagation of optical wave fields at any distance from the aperture the Fresnel–Kirchhoff diffraction integral must be calculated. The numerical implementation of this integral is very complex computationally, thus preventing any video-rate application that uses it. To surpass this problem Angular spectrum, Fresnel–Bluestein and Fresnel approaches have been implemented. Angular spectrum is used in any optical setup with small propagating distances. Fresnel is the fastest implementation for large propagation distances but without control of the resulting scaling of the propagated wavefields. Fresnel–Bluestein eliminates this latter problem with a slightly higher computational complexity. Solution method: Angular spectrum, Fresnel–Bluestein and Fresnel approaches for the numerical propagation of optical fields in the JAVA computation environment. Additional comments: Available for download from URL: http://unal-optodigital.github.io/JDiffraction/. The API can be found here: http://unal-optodigital.github.io/JDiffraction/javadoc/index.html References:[1]P. Piedrahita-Quintero, C. Trujillo, J. García-Sucerqui, JDiffracto 1.2 API, (2016).http://unaloptodigital.github.io/JDiffraction/javadoc/index.html (accessed June 7, 2016).