Heidelberg 2022 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

T: Fachverband Teilchenphysik

T 69: DAQ and Trigger 3

T 69.1: Talk

Wednesday, March 23, 2022, 16:15–16:30, T-H28

Development of an FPGA Implementation of Convolutional Neural Networks for Signal Processing for the Liquid-Argon Calorimeter at ATLAS — Anne-Sophie Berthold, •Nick Fritzsche, Markus Helbig, Rainer Hentges, Arno Straessner, and Johann Christoph Voigt — Institut für Kern- und Teilchenphysik (IKTP), TU Dresden, Germany

The Phase-II upgrade of the ATLAS detector will prepare for the high-luminosity phase of the LHC, where the number of proton-proton collisions occurring at the same time will increase significantly. This leads to higher requirements for the data processing, since the rate of detected particles in one detector cell will increase. New machine learning solutions are under development to better reconstruct the energy deposited in the calorimeter and its timing information than the current optimal filter approach.

This talk introduces the implementation of convolutional neural networks for FPGA hardware. The application of time division multiplexing is discussed, which is necessary to cover the high number of detector readout channels and to reuse the network for multiple input streams. The latest performance results in terms of FPGA resource usage, achievable operation frequency and latency are presented. To verify the hardware implementation, a software reference model was created and the precision of the calculation results was analyzed. At last, first preparations for tests on hardware are shown.