Keyword: data-analysis
Paper Title Other Keywords Page
WEPG07 A Heterogeneous FPGA/GPU Architecture for Real-Time Data Analysis and Fast Feedback Systems GPU, FPGA, monitoring, detector 626
  • M. Vogelgesang, L.E. Ardila Perez, M. Caselle, S.A. Chilingaryan, A. Kopmann, L. Rota, M. Weber
    KIT, Eggenstein-Leopoldshafen, Germany
  We propose a versatile and modular approach for a real-time data acquisition and evaluation system used for monitoring and feedback control in beam diagnostic and photon science experiments. Our hybrid architecture is based on an FPGA readout card* and a GPU for data processing. To increase throughput, lower latencies and reduce overall system strain, the FPGA write data directly in the GPU. After real-time data analysis the GPU writes back results either directly to the FPGA in case of fast feedback systems or to the CPU host system for storage. Communication and scheduling are handled transparently by our processing framework**. However, users can customize and extend it with their own processing plugins. Although the system is designed for real-time purposes, the modular approach also allows standalone usage for high-speed off-line analysis. We evaluated the performance of our solution measuring both processing times of data analysis algorithms used with beam instrumentation detectors as well as transfer times between FPGA and GPU. The latter suggests throughputs of up to 6.5 GB/s with latencies down to tens of microseconds, thus making it suitable for fast feedback systems.
* A PCIe DMA Architecture for Multi-Gigabyte Per Second Data Transmission, 10.1109/TNS.2015.2426877
** A Scalable GPU-based Image Processing Framework for On-line Monitoring, 10.1109/HPCC.2012.116
poster icon Poster WEPG07 [17.144 MB]  
DOI • reference for this paper ※ DOI:10.18429/JACoW-IBIC2016-WEPG07  
Export • reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)