Keyword: GPU
Paper Title Other Keywords Page
WEPSB025 Application of GPGPUs and Multicore CPUs in Optimization of Some of the MpdRoot Codes framework, distributed, detector, experiment 416
 
  • A.I. Fatkina, O.O. Iakushkin, N.O. Tikhonov
    Saint Petersburg State University, Saint Petersburg, Russia
 
  Funding: This research was partially supported by SPbU (Saint Petersburg State University) grants 9.37.157.2014, 0.37.155.2014 and Russian Foundation for Basic Research grant (project no. 16-07-01113).
We analyzed the ways to optimize MPDRoot algorithms using existing solutions from external libraries. We also examined the libraries designed to work with graphics accelerators and multi-core CPUs, such as cuRAND, cuFFT and OpenCL FFT. The paper describes the ways to expedite a portion of Kalman filter by transferring it to GPUs or multi-core CPUs using the implementation included into the MPDRoot package.
 
Export • reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)  
 
WEPSB030 Parallelization of Envelope Dynamics of High Intensive Beams space-charge, simulation, software, controls 425
 
  • N.V. Kulabukhova
    St. Petersburg State University, St. Petersburg, Russia
 
  In this work the survey of methods for high intensive beam dynamics is given. As an alternative to them the approach based on envelope dynamics was used. This method is focusing on the use of the matrix form for Lie algebraic methods for calculating the beam dynamics in the presence of self-field of the beam. In particular, the corresponding calculations are based on the predictor-corrector method. Pros and cons of using described approach on hybrid systems are discussed.  
Export • reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)  
 
WEPSB051 Mathematical and Computer Methods of Data Processing in Nuclear Medicine Studies software, detector, photon, experiment 480
 
  • E.D. Kotina, A.V. Babin, P. V. Bazhanov, D.A. Ovsyannikov, V.A. Ploskikh, Yu. Shirokolobov
    Saint Petersburg State University, Saint Petersburg, Russia
 
  Currently nuclear medicine is a high-tech field. Its development requires solutions of problems related both to the improvement of hardware and computer processing of the information obtained in the course of study. The basic types of hardware of nuclear medicine are gamma cameras and single photon emission computed tomography (SPECT), positron emission tomography (PET) and hybrid scanners (SPECT/CT, PET/ CT). The methods of SPECT data processing include analysis of static, dynamic, tomographic and ECG-gated images. The PET data reconstruction software restores the three-dimensional distribution of the radiopharmaceutical in the body. Performance of iterative methods increases with the use of modern graphics processors. Three-dimensional imaging allows a detailed analysis of the study area. The software also implements a fusion imaging of SPECT/CT, PET/CT scans performed in the same coordinate system. An application of data flow model in medical software development is considered. Web-based imaging front-end of storage and processing system is presented.  
Export • reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)