PHYSICS PERFORMANCE AND DATASET (PPD)

 

Introduction

Since July, the activities have been focused on very diverse subjects: operations activities for the 2012 data-taking, Monte Carlo production and data re-processing plans for 2013 conferences (winter and summer), preparation for the Upgrades TDRs and readiness after LS1.

The regular operations activities have included: changing to the 53X release at the Tier-0, regular calibrations updates, and data certification to guarantee certified data for analysis with the shortest delay from data taking. The samples, simulated at 8 TeV, have been re-reconstructed using 53X. A lot of effort has been put in their prioritisation to ensure that the samples needed for HCP and future conferences are produced on time.

Given the large amount of data that have been collected in 2012 and the available computing resources, a careful planning is needed. The PPD and Physics groups worked on a master schedule for the Monte Carlo production, new conditions validation and data reprocessing.

The CMSSW release 53X will be used for future conferences 2013. No general reprocessing of the collision data will be done for the winter conferences, but only few specific primary datasets for the Higgs analyses will be reprocessed with improved ECAL calibration for 2012 C and D. For summer 2013, a complete reprocessing of the data will be performed including all the latest conditions and calibrations.

The preparation of the various TDRs for the Upgrade projects required the production of specific Monte Carlo samples. Given the increasing attention on physics studies for HL-LHC, in the past few months, the PdmV effort has been on integrating the existing Physics validation tools and procedures.

The Global Event Description team had a three days workshop at FNAL in August that will be summarised in a detailed plan to be made in collaboration with the POGs and Reconstruction coordinators and presented at the December CMS Week. This plan will describe the steps for the reconstruction improvements from now to September 2013.

Finally, an overall schedule of the PPD activities for LS1 will be set-up during the PPD Workshop between 28 and 30 November.

Alignment and Calibration and Database (AlCaDB)

Most of the time of the AlCaDB project concentrated on supporting the data-taking and data-analysis activities of CMS. In between that, work was on-going on new tools to ease that work and make it more robust. On the AlCa side, a new web tool to manage the Global Tags was introduced in the summer, and a new and significantly improved version of the DropBox tool to upload conditions to the DB was developed. Both tools are now in use and their usage will ease the work of the AlCaDB team and the users for the upcoming data-taking as well as the future MC productions and data/MC reprocessings.

Data Quality Monitoring (DQM)/Data Certification

DQM has been working smoothly in online and offline environments. Data Certification is carried out regularly and the certification efficiency is steadily increasing, thanks to the effort of all parties involved, from Run Coordination to shifters and certification experts. Yet times are changing, and we need to shift our attention to the LS1 development period. We will have to face many interesting challenges, from multicore, multithread event processing in CMSSW to an improved, but radically changed, DAQ processing. We need to be able to profit from this exciting period to rethink the main DQM Framework and eventually make it an even better and more flexible system. One of our main goal here is to establish a strong DQM development team and hopefully build a solid and spread DQM culture among the collaboration: your contribution and participation in this effort is vital for our success!

Physics Data Monte-Carlo Validation (PdmV)

A validation book-keeping system for CMS (valDB) has been commissioned and fully deployed for centralisation of the validation reports from more than 80 collaborators scrutinising the effect of changes on software, conditions and trigger selection. The effort of joining the upgrade activity in the mainstream of validation is well under way. New prompt workflows have been deployed to validate the changes of calibrations needed in the HLT and in the prompt reconstruction to compensate for radiation-induced aging effects. We are planning development for LS1, at which time collaborators will shift their activity towards development. Several monitoring services and scheduling tools for Monte Carlo production have been put in place to support the complex needs from analysis community, merging experience from Computing, PPD, and Offline. The re-writing of the production management system (formerly known as PREP) is well under-way and will be ready for deployment during the winter. We are looking forward to working on other aspects of Data and Monte-Carlo validation during the long shutdown, in preparation for 2013 summer conferences and for ever-improved quality of CMS data.


by L. Silvestris, P. Azzi, C. Cerminara, with contributions from R. Castello, M. De Mattia, A. Pfeiffer, C. Rovelli, F. De Guio, S. Dutta, K. Maeshima, M. Rovere, G. Boudoul, G. Franzoni and J.-R. Vlimant. Edited by K. Aspola.