Detector integration at HEPS: a systematic, efficient and high-performance approach

Qun Zhang 1,2, Peng-Cheng Li 1,3,4, Ling-Zhu Bian 1, Chun Li 1 · 2024

Read Detector integration at HEPS: a systematic, efficient and high-performance approach by Qun Zhang 1,2, Peng-Cheng Li 1,3,4, Ling-Zhu Bian 1, Chun Li 1 (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2411-01260.txt.

Description

Abstract 1 1 footnotetext: Correspondence e-mail: [email protected] . 1 1 footnotetext: Institute of High Energy Physics, Chinese Academy of Sciences, Beijing 100049, People’s Republic of China. 2 2 footnotetext: North China University of Technology, Beijing 100144, People’s Republic of China.

Catalog record

CREATORQun Zhang 1,2, Peng-Cheng Li 1,3,4, Ling-Zhu Bian 1, Chun Li 1
YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE35 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2411-01260

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Detector integration at HEPS: a systematic, efficient and high-performance approach Qun Zhang 1,2, Peng-Cheng Li 1,3,4, Ling-Zhu Bian 1, Chun Li 1 Abstract 1 1 footnotetext: Correspondence e-mail: [email protected] . 1 1 footnotetext: Institute of High Energy Physics, Chinese Academy of Sciences, Beijing 100049, People’s Republic of China. 2 2 footnotetext: North China University of Technology, Beijing 100144, People’s Republic of China. 3 3 footnotetext: National Synchrotron Radiation Laboratory, University of Science and Technology of China, Hefei, Anhui 230029, People’s Republic of China. 4 4 footnotetext: University of Chinese Academy of Sciences, Beijing 100049, People’s Republic of China. Synopsis Keywords: detector integration, GenICam, high-performance readout, Python IOC, software architecture. To integrate the at least 25 kinds of detector-like devices to be used at HEPS Phase I, a systematic and efficient strategy is applied. A careful separation of concerns between collaborating groups of personnel is followed to boost efficiency, and an extended ADGenICam is used to reduce code repetition in EPICS IOCS for detectors. An areaDetector workalike in the QueueIOC Python IOC framework is used to integrate high-performance detectors. Abstract At least 25 kinds of detector-like devices need to be integrated in Phase I of the High Energy Photon Source (HEPS), and the work needs to be carefully planned to maximise productivity with highly limited human resources. After a systematic analysis on the actual work involved in detector integration, a separation of concerns between collaborating groups of personnel is established to minimise the duplication of efforts. To facilitate software development for detector integration, the ADGenICam library, which abstracts repeated code in EPICS modules for cameras, is extended to support a much wider range of detectors. An increasingly considerable fraction of detectors, both inside and outside HEPS, offer performance that exceed capabilities of the areaDetector framework in EPICS . Given this background, areaDetector ’s limitations in performance and architecture are analysed, and a QueueIOC -based framework that overcomes these limitations is introduced. A simple, flexible ZeroMQ -based protocol is used for data transport in this framework, while RDMA transport and multi-node readout will be explored for higher data throughputs. By calling C/C++ libraries from within Python, the performance of the former and the expressiveness of the latter can coexist nicely; the expressiveness allows for much higher efficiency in the implementation and use of integration modules functionally comparable to their EPICS counterparts. 1 Introduction In recent years, progress is being rapidly made in the research and construction of advanced light sources, eg. 4th-generation synchrotrons and free-electron lasers. While pushing towards the physical limits, the excellent attributes of these light sources – small light spots, high brightness, strong coherence etc – allow scientists and engineers to do researches in much finer spatial and temporal scales. Correspondingly, radiation detection techniques ( Fajardo 2017 ) are also advancing rapidly, providing instrumental bases for practical exploitation of the attributes above. Improvements on the pixel sizes (in combination with the numbers of pixels) and framerates of detectors enable exploration of matter on finer spatial and temporal scales, respectively. Higher framerates also allow for the use of techniques like fly scans ( Li et al. 2023b ; Li et al. 2023a ) , which greatly reduce the time spent on scanning which increases because of the smaller light spots and the more scan points; these techniques also help to reduce the radiation damage in samples because of the higher brightness. Meanwhile, we have noticed that the mainstream software frameworks used by major light sources around the world may have themselves become a main bottleneck in the application of high-performance detectors. Both the Beijing Synchrotron Radiation Facility (BSRF) and the High Energy Photon Source ( Xu et al. 2023 , HEPS, cf. ) use the Experimental Physics and Industrial Control System ( EPICS ) as the main environment for device control; the main framework for EPICS -based detector integration is areaDetector ( Rivers 2017 ) , which is the main example in our discussion. According to our experience, when writing HDF5 files using areaDetector ’s HDF5Plugin , the data throughput has an upper bound of around 500-600 MB/s. In contrast, even only among the requirements of detector integration from the 15 beamlines in Phase I of HEPS (Table 1 ), there are already quite a few detectors exceeding 1 GB/s (some even exceeding 10 GB/s) throughput, each with some beamline(s) wanting to exploit its performance as much as reasonable. Outside BSRF and HEPS, readout requirements reaching 1 GB/s or even 10 GB/s are also becoming increasingly common ( Fischer et al. 2017 ; Yendell et al. 2017 ; Göries et al. 2023 ; Dong et al. 2022 ) . Table 1: A few representative high-performance detectors that need to be integrated at HEPS. Detector model Max. resolution Max. framerate Max. throughput Tucsen camera A 6 K \times 6 K \times 2 B 19 Hz 1.3 GB/s Tucsen camera B * 5 K \times 4 K \times 1 B 450 Hz 8.8 GB/s A self-made Si-pixel detector 2 \times 5.76 M \times 2 B † 1000 Hz 22 GB/s Eiger X 1M 1 K \times 1 K 3000 Hz Maybe > 1 GB/s ‡ Hamamatsu C15550-20UP 4096 \times 2304 \times 2 B 120 Hz 2.1 GB/s * Still in production. † With dual thresholds. ‡ (Even) after compression. Apart from the data throughput issue above, we have also found areaDetector HDF5Plugin has an about 4 kHz upper bound in framerates when writing HDF5 files; meanwhile, fly-scan experiments on the 10 kHz level ( Deng et al. 2019 ; Batey, Rau and Cipiccia 2022 ) are a field pursued actively by us. Even only among the requirements from HEPS Phase I, experiments exceeding 10 kHz at least include coherent diffraction imaging (CDI) and X-ray photon correlation spectroscopy (XPCS) experiments at the hard X-ray coherent scattering beamline (B4), as well as 2D real-space current mapping experiments at the high-resolution nanoscale electronic structure spectroscopy beamline (BC). Apart from possibly demanding high-throughput readout, these experiments may also demand data readout from sensors, motion controllers, fly-scan controllers ( Christian et al. 2019 , like PandABox, cf. ) etc . Although the latter kind of readout has high framerates, its throughputs may not be very high; because of the smaller size of each data frame, the devices that produce these data may be regarded as 0D/1D detectors ( Li et al. 2023a ) . Unfortunately, although areaDetector supports 1D detectors like Mythen, it does not seem to consider 0D readout at all, making it non-trivial to integrate 0D devices in a unified way. Aside from the performance considerations above, another issue we notice is the costs of integration, where the first problem is the number of detector brands: even only at HEPS Phase I, we have at least 25 kinds of detector-like devices to be integrated (Table 2 ), and we believe similar problems will also occur at other light sources, especially new light sources. Only about a half among these devices have open-source EPICS modules (input/output controllers or IOCs), and in our experience a majority of these modules also need some customisations or even fixes. Among the other half, most require us to write high-performance integration modules, while the rest need self-developed EPICS IOCs. During the

Read the complete text