How the Monte Carlo production of a wide variety of different samples is centrally handled in the LHCb experiment

21st International Conference on Computing in High Energy and Nuclear Physics

物理学;计算机科学

Corti, G.^1 ; Charpentier, Ph^1 ; Clemencic, M.^1 ; Closier, J.^1 ; Couturier, B.^1 ; Kreps, M.^2 ; Màthè, Z.^1 ; Hanlon, D.O.^2 ; Robbe, P.^1,3 ; Romanovsky, V.^4 ; Stagni, F.^1 ; Zhelezov, A.^5

CERN, Geneva, Switzerland^1

Department of Physics, University of Warwick, Coventry, United Kingdom^2

LAL, Universite Paris-Sud, CNRS/IN2P3, Orsay, France^3

Institute for High Energy Physics (IHEP), Protvino, Russia^4

Physikalisches Institut, Ruprecht-Karls-Universität Heidelberg, Heidelberg, Germany^5

关键词: Common infrastructures;    Distributed systems;    Event Types;    Identification codes;    Level of automations;    Physics programs;    Production environments;    Specific tasks;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/664/7/072014/pdf
DOI  :  10.1088/1742-6596/664/7/072014

学科分类：计算机科学（综合）

来源: IOP

【 摘 要 】

In the LHCb experiment a wide variety of Monte Carlo simulated samples needs to be produced for the experiment's physics program. Monte Carlo productions are handled centrally similarly to all massive processing of data in the experiment. In order to cope with the large set of different types of simulation samples, necessary procedures based on common infrastructures have been set up with a numerical event type identification code used throughout. The various elements in the procedure, from writing a configuration for an event type to deploying them on the production environment, from submitting and processing a request to retrieving the sample produced as well as the conventions established to allow their interplay will be described. The choices made have allowed a high level of automation of Monte Carlo productions that are handled centrally in a transparent way with experts concentrating on their specific tasks. As a result the massive Monte Carlo production of the experiment is efficiently processed on a world-wide distributed system with minimal manpower.