Observatorio de I+D+i UPM

Memorias de investigación
Communications at congresses:
High Performance Image Acquisition and Processing Architecture for Fast Plant System Controllers based on FPGA and GPU
Year:2015
Research Areas
  • Electronic technology and of the communications
Information
Abstract
Real time image processing is an area of increasing interest to the nuclear fusion community as the number of cameras present in large fusion devices, both in the visible and infra-red ranges, increases continuously [1]. The amount of data generated by image diagnostics requires including real time processing capabilities in the acquisition process in order to make them effective tools. At least there is a need to include some image pre-processing in the acquisition process, such as image compression. But as we advance towards long pulse devices, there is an increasing need to include more advanced functionalities, such as pattern recognition, or event detection. The two dominant technologies that are being used in this field are FPGA and GPU due to their algorithm parallelization capabilities. But not much work has been done to standardize how these technologies can be integrated in data acquisition systems where control and supervisory requirements are in place, such as ITER. This work tests an Image Acquisition and processing system for Cameralink devices based in a FPGA DAQ device (National Instruments FlexRIO technology) and a NVIDIA Tesla GPU series card. This integration, compliant with ITER fast controller solutions, simplifies the implementation of advanced data (images) acquisition systems. The most important advantage of this solution is that data acquired from the set NI1483-NIPXIe7966R is moved directly to GPU avoiding the use of CPU memory buffers using NVIDIA GpuDIRECT RDMA [2] technology. Thus it?s possible to increase the performance of the system because CPU intervention is minimized because data transfer is done using PCI Express and DMA. The system has been developed using CODAC Core System standards (EPICS and Nominal device Support [3]) in order to guarantee an easy integration in ITER?s solutions.
International
Si
Congress
10th IAEA Technical Meeting on Control, Data Acquisition and Remote Participation for Fusion Research
960
Place
Ahmedabad
Reviewers
Si
ISBN/ISSN
00000000
Start Date
20/04/2015
End Date
24/04/2015
From page
1
To page
5
10th IAEA Technical Meeting on Control, Data Acquisition and Remote Participation for Fusion Research
Participants
  • Autor: Julian Nieto Valhondo (UPM)
  • Autor: Guillermo de Arcas Castro (UPM)
  • Autor: Mariano Ruiz Gonzalez (UPM)
  • Autor: Jesus Vega (CIEMAT)
  • : Rodrigo Castro
Research Group, Departaments and Institutes related
  • Creador: Grupo de Investigación: Investigación en Instrumentación y Acústica Aplicada (I2A2)
  • Departamento: Ingeniería Telemática y Electrónica
S2i 2020 Observatorio de investigación @ UPM con la colaboración del Consejo Social UPM
Cofinanciación del MINECO en el marco del Programa INNCIDE 2011 (OTR-2011-0236)
Cofinanciación del MINECO en el marco del Programa INNPACTO (IPT-020000-2010-22)