Abstract
In 2015, CMS was the first LHC experiment to begin using a multi-threaded framework for doing event processing. This new framework utilizes Intel’s Thread Building Block library to manage concurrency via a task based processing model. During the 2015 LHC run period, CMS only ran reconstruction jobs using multiple threads because only those jobs were sufficiently thread efficient. Recent work now allows simulation and digitization to be thread efficient. In addition, during 2015 the multi-threaded framework could run events in parallel but could only use one thread per event. Work done in 2016 now allows multiple threads to be used while processing one event. In this presentation we will show how these recent changes have improved CMS’s overall threading and memory efficiency and we will discuss work to be done to further increase those efficiencies.
Highlights
Since the beginning of the LHC Run 2 era, CMS has been using multiple threads in the offline data processing framework [1] when doing production level work
The online high level trigger farm switched to using multiple threads per job in September of 2015
Looking at the times when a red module has stalled we can see they correspond to times where another stream is using multiple threads
Summary
To cite this article: C D Jones and CMS Collaboration 2017 J. View the article online for updates and enhancements. - Multi-threaded Object Streaming Salvatore Di Guida, Giacomo Govi, Miguel Ojeda et al. - Implementation of a Multi-threaded Framework for Large- scale Scientific Applications E Sexton-Kennedy, Patrick Gartung, C D Jones et al. - Recent Standard Model results from CMS Simon de Visscher and CMS collaboration. C D Jones on behalf of the CMS Collaboration Fermilab, P.O.Box 500, Batavia, IL 60510-5011, USA
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
