Gießen 2024 – scientific programme

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

HK: Fachverband Physik der Hadronen und Kerne

HK 54: Computing II

HK 54.6: Talk

Wednesday, March 13, 2024, 18:45–19:00, HBR 19: C 103

Data challenges at CBM - towards scalable workflows — •Andreas Redelbach for the CBM collaboration — Frankfurt Institute for Advanced Studies, Goethe University, Frankfurt, Germany

Operating the CBM experiment at interaction rates up to 10 MHz requires data reduction in real-time. This necessitates highly efficient online processing of measurements and the underlying algorithms. More specifically, since the free-streaming readout data are processed in software, the performance of the reconstruction algorithms is a critical issue. A promising option for acceleration is based on an efficient parallelization of data processing developed for both CPU and GPU architectures. A number of measures have been taken to minimize runtimes of reconstruction algorithms and to optimize the scaling of some time-critical workflows.

Using the mCBM full-system test setup at SIS18 allows testing of all relevant components connected to study all processing steps. It is interesting to note that progress has been achieved in particular using test data from mCBM beamtimes. In this contribution, some of the concepts and recent progress towards high throughput processing in the CBM reconstruction chain are summarized.

This work is supported by BMBF (05P21RFFC1).

Keywords: CBM; Online Computing; HLT