Preparing CMS for the HL-LHC and the future of computing

Preparing CMS for the HL-LHC and the future of computing
Allison Hall, Fermilab
Date and time: Fri, Apr 16, 2021 - 2:15pm
Location: Zoom only
Category: ACFI Seminar

The High Energy Physics (HEP) community is approaching the High-Luminosity LHC (HL-LHC) upgrade, where the instantaneous luminosity will increase by a factor of 10 relative to the LHC design. This will provide exciting physics opportunities but also unprecedented computing challenges for the CMS and ATLAS experiments. In this talk, I will cover two innovative R&D efforts designed to help reduce the computing requirements and maximize the discovery potential of the HL-LHC. The first initiative is the Compact Framework for Effective Analysis (Coffea), the first fully functioning HEP analysis framework based on the usage of Big Data technologies. The second project is the mkFit algorithm, a highly parallelized version of the traditional Kalman filter algorithm for charged particle tracking. Through the efficient use of modern CPU architectures, mkFit has achieved a factor of 6 speedup with respect to the nominal CMS algorithm, with comparable physics performance. Lessons learned and future opportunities and plans will be discussed.