Comparison of distributed DAQ software frameworks for the high-throughput TUPI detector

Timepix-based Ultra-fast Photon Imaging (TUPI) is a photon-counting hybrid detector family proposed to fulfill the requirements of ORION's (Brazil's planned BSL-4 laboratory) tender and hard X-rays beamlines. The first detector model is planned to use an arrangement of 3x3 Timepix4 ASICs to provide 1344x1536 pixel images (55 μm pixel pitch) at 16-bit dynamic range, with expected acquisition rates up to 4kHz in frame-based mode. However, the hardware can reach up to 44kHz, and other chip arrangements are possible in the future, so a future-proof software solution is being sought after among currently available open-source Data Acquisition (DAQ) software frameworks, such as LImA2, Odin and AreaDetector. A distributed DAQ system could simplify server hardware requirements, given that a 3x3 ASIC array acquiring at maximum frame rate requires 144 dedicated 10.24Gbps links, which can be spread across as many hardware units as needed. Furthermore, the decentralized software architecture can help in performance sensitive tasks like file-saving, enabling it to keep up with the detector rates. This work aims to compare available frameworks considering multiple aspects, such as performance, ergonomics (i.e. developer experience and deployment procedure) and control system integration (EPICS, in our case). The suitability of the frameworks for online processing and distributed acquisition will also be compared, as current and future challenges might make these features compulsory.