dscout

40 MHz datascouting analysis framework (CMS@CERN)

This project is about …

analyzing the first-ever-recorded CMS@CERN scouting data at 40MHz(!) and searching for regularities and irregularities therewithin. The rate, directionality, charge and possible time dependences of the recorded cosmic/collision muons will be studied. The size of the sample could easily become unmanageable (big data) if not smart slicing and selection is applied. Industrial standards will be used for the analysis exploiting modern libraries used in big data/machine learning.

Introduction

Searching for the Higgs boson and beyond the standard model phenomena at the LHC demands many proton-proton (pp) collision events at the highest possible achievable energy (13 TeV).

By design, the LHC produces far too many collisions to be recorded on disk. With an LHC bunch crossing (BX) rate of 40MHz (1 BX per 25ns) and using the latest greatest technologies (FPGAs + computing farms) available upon its construction, the CMS experiment can “briefly look” at all BX’s, but preserve (i.e., trigger) for later analysis only a tiny fraction of them, ~2.5 out of every million (10^6) of BX’s. Certain rules have to be applied in order to prevent CMS from attempting to record too many pp collisions in short time, namely no more than 1 trigger-accept in 3 BX’s, 2 in 25, 3 in 100, 4 in 240.

At least these all were true once upon a time. In practice, CMS collaboration managed to improve the situation by a factor of ~2, in the late times of the Run2. Still, there is no waterproof assurance that we don’t miss from seeing interesting events that are produced during the pp collisions but the trigger (event selection) logic has not been programmed to be aware and anticipate for. In particular, slowly evolving phenomena that extend in time in more than 1 BX (>25 ns), would be nearly next to impossible to trigger on with the present architecture. In addition, final states in regions of the phase space where the background is large are technically limited by the available bandwidth of the trigger.

The 40 MHz scouting, is a novel CMS R&D project, having as ambition to bypass the present architectural limitations enabling the recording of ALL of the collision events without the need of a trigger-accept, at a price of reduced resolution on the recorded “pictures”. In a pilot run during 2018, the 40 MHz scouting system recorded for first time collision and cosmic rays data. In this first test-run, only the Muon subsystem of the CMS detector participated in the scouting. The full blown production-level system is envisioned to be ready for phase-II (>2027). Further pilot test-runs are scheduled for Run 3.

Plan

Download example files with cosmic rays

Example data files from run 326790 can be downloaded here.

Analysis

The analysis will be carried out in two steps:

Keywords