Abstract
The hardware landscape used in HEP and NP is changing from homogeneous multi-core systems towards heterogeneous systems with many different computing units, each with their own characteristics. To achieve maximum performance with data processing, the main challenge is to place the right computing on the right hardware. In this paper, we discuss CLAS12 charge particle tracking workflow orchestration that allows us to utilize both CPU and GPU to improve the performance. The tracking application algorithm was decomposed into micro-services that are deployed on CPU and GPU processing units, where the best features of both are intelligently combined to achieve maximum performance. In this heterogeneous environment, CLARA aims to match the requirements of each micro-service to the strength of a CPU or a GPU architecture. A predefined execution of a micro-service on a CPU or a GPU may not be the most optimal solution due to the streaming data-quantum size and the data-quantum transfer latency between CPU and GPU. So, the CLARA workflow orchestrator is designed to dynamically assign micro-service execution to a CPU or a GPU, based on the online benchmark results analyzed for a period of real-time data-processing.
Highlights
We are witnessing unprecedented expansion of scientific data
An efficient operation on these platforms will require a tight integration of components handling the data flow, distribution and processing, as well as workflow management
It is important to mention that the workflow orchestrator is not interfering with streaming data-processing and assumes a passive observer role during the data processing, monitoring, and performing a real-time benchmarking for each microservice
Summary
We are witnessing unprecedented expansion of scientific data This is explained by upgrades to existing accelerator facilities, such as LHC and CEBAF, as well as improvements in detector equipment. It is expected that the traditional software and hardware infrastructures used in HEP and NP will have difficulties processing large amount of data generated by new experiments [1]. Data-processing application development frameworks must be able to integrate data processing components that internally use different parallelization models, such as physics generators and detector simulation. They must encourage common development and code reuse by creating methods to integrate new developments with decades of legacy work across experiments
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
