Simulation of the 4 Layer Pixel Upgrade System

Simulation of the 4 Layer Pixel Upgrade System Harry Cheung (Fermilab) (For the Tracker Upgrade Simulations Group) Outline Technical/Software Status Past Work on Phase 1 Simulation Studies Current Phase 1 Simulation Studies Future Plan and Task List See the talks in the Wednesday 14: 00 -16: 00 Tracker Upgrade Simulations session; and Tuesday 14: 00 -18: 00 Track Trigger session for the latest results of simulation studies

Technical/Software: Geometry Full and Fast Simulations Geometry We have a Phase 1 geometry setup for Full and Fast Simulations • BPIX ~ from Roland’s drawings (implemented by X. Huang, H. Cheung) FPIX ~ from Simon’s drawings (implemented by P. Jindal) – Includes option for smaller radius inner barrel layer (and beam pipe) – Configurable sized pixels in each layer Std Geom Phase 1 Pixel Upgrade Phase 1 Meeting, 27 April 2010 H. W. K. Cheung (FNAL) 2

Technical/Software: Geometry Rechits from simulations of dimuons: Smaller radius inner layer option Can change pixel size via a single XML geometry file for each layer and disk Pixel Upgrade Phase 1 Meeting, 27 April 2010 H. W. K. Cheung (FNAL) 3

Technical/Software: Geometry Green Histo: Phase 1 Dots: Std geometry CMSSW_3_3_6 Material budget: Still some issues BPIX cooling reweigh only, using std. geom. ladder and module construction FPIX (services) material not completely implemented Material budget plots from Paul Sheldon Pixel Upgrade Phase 1 Meeting, 27 April 2010 H. W. K. Cheung (FNAL) 4

Technical/SW: Fast Simulation with digis for faster geometry variation studies Fast. Sim Full. Sim Fast Simulation “interaction geometry” also done and tuned to Full Sim However additional issues with the Fast simulation Illustration of Fast Simulation “interaction geometry”. Standard geometry is shown. Pixel Upgrade Phase 1 Meeting, 27 April 2010 H. W. K. Cheung (FNAL) 5

Technical/SW: Fast Simulation Green Histo: Fastsim Dots: Fullsim CMSSW_3_3_6 Material budget: Matched to Full Simulation Code to directly compare material budget between Fast and Full simulations Phase 1 Std Material Geom budget plots from Paul Sheldon Pixel Upgrade Phase 1 Meeting, 27 April 2010 H. W. K. Cheung (FNAL) 6

Technical/SW: Fast Simulation additional issue: Tracker occupancies No out-of-time pileup for Fast Simulation “In-time” Tracker occupancies differ in Fast and Full Sim: still investigating • Illustration: Rates from first “Stack Layer” in Long barrel geometry: Occupancy Simhit Digi ratesplots from Emmanuele Salvati (rates for min-bias and 50 PU x 4 /module) Pixel Upgrade Phase 1 Meeting, 27 April 2010 H. W. K. Cheung (FNAL) 7

Technical/SW: Hits & Tracking efficiency for dimuons with non-iterative (CMSSW_1_8_4 -like) tracking and triplet seeds: Compare smaller radius inner layer option Pixel Upgrade Phase 1 Meeting, 27 April 2010 H. W. K. Cheung (FNAL) 8

Technical/SW: Hits & Tracking efficiency for dimuons with non-iterative tracking: Still issues for Fast simulation matching Full Sim in new FPIX Std Geom Phase 1 Pixel Upgrade Phase 1 Meeting, 27 April 2010 H. W. K. Cheung (FNAL) 9

Technical/SW: Hits & Tracking Still some technical work/issues with hits & tracking software Rechit (cluster) position and errors • Pixel hit positions and errors from code optimized for standard geometry • Need to create appropriate code for upgrade geometry – Account for variation of pixel sizes – Account for new FPIX construction Out-of-memory problems at high pileup (CMSSW_3_3_6) • Full simulation crashes at about 2 1034 cm-2 s -1 (cannot fix) • Tracking in Fast simulation crashes at about 4 1034 cm-2 s -1 due to limit on number of clusters in code (can fix) Tuning of iterative tracking steps • Steps 0 -4 changed to use triplet seeds instead of pair seeds – Quartet seeds not yet implemented in code • Steps 5 -6 containing strip pair seeds has to be removed for high PU • Need to tune the selection/seeding for all steps Pixel Upgrade Phase 1 Meeting, 27 April 2010 H. W. K. Cheung (FNAL) 10

Technical/Software Summary Full Simulation Ready to use for Phase 1 simulation studies: • Includes smaller radius inner layer option, and configurable pixel sizes Some improvements and geometry variation needed: • Improve BPIX material description; Put in FPIX services • Need geometry (variation) with 4 th layer at larger radius Fast Simulation Ready for Phase 1 simulations studies but take into account: • Need understand small Fast/Full sim differences at high eta • Need to understand tune tracker/pixel occupancies • Implement out-of-time pileup (if needed? ) Hits and Tracking Fully working, but accurate results still need some work: • Need hit position and error routine for upgrade geometry • Need to tune iterative tracking and implement quartet seeds Pixel Upgrade Phase 1 Meeting, 27 April 2010 H. W. K. Cheung (FNAL) 11

Past Simulation Studies Much of the upgrade simulations efforts in the past were on Phase 2 Past Phase 1 simulation studies is of less interest now Used “Quick Approach” Phase 1 Geometry • 4 layer upgrade BPIX but still with half and full ladders, though material was weighted for new lighter services • 3 FPIX disk (but same FPIX disks as standard geometry) Later used some combination of “Quick Approach” and early version of “Detailed Approach” Phase 1 Geometry • Problems in FPIX region; BPIX layers not finalized Many studies used CMSSW_1_8_4 -like non-iterative tracking • Comparisons with the standard geometry should be will the most recent (possible) CMSSW software - at least the iterative tracking – Also iterative tracking easier for tuning to upgrade conditions • Many other improvements to the code in CMSSW_3_3_6 (e. g. memory usage) and bug fixes (e. g. truth particle parentage) Next 5 slides shown just as examples of most recent past Phase 1 studies • From Feb 11 Status report given by Pratima Jindal (Tracker Week) Pixel Upgrade Phase 1 Meeting, 27 April 2010 H. W. K. Cheung (FNAL) 12

From Feb 11 Status Report: Tracking Efficiency and Fakerate Studies Fast Simulation (with digis) using CMSSW_2_2_6 Phase 1 geometry using realistic Phase 1 disks with problems in negative eta and the quick approach for adding the 4 th barrel layer (version from Oct), will update soon with new Higher fake rate geometry description for Phase 1 QCD dijet events p. T= 100 -120 Ge. V, |η| < 2. 5 for PU 100 Track reconstruction: • Using old non-iterative tracking • Standard geometry: • triplet seeds (3 -out-of-3) • Phase 1: • triplet seeds (3 -out-of-4) • Minimum of 3 hits per track • Restricted vertex search volume in Tracking. Truth. Producer Requiring 6 hits per track reduces fake rate for PU 100 John Ellison & Ferdinando Giordano Pixel Upgrade Phase 1 Meeting, 27 April 2010 H. W. K. Cheung (FNAL) 13

From Feb 11 Status Report: b-tagging Performance Using Phase 1 geometry using realistic Phase 1 disks with problems in negative eta and the quick approach for adding the 4 th barrel layer (version from Oct) • Using 50 – 120 Ge. V jets • Fastsim using CMSSW_2_2_6 • Track Counting High Efficiency algorithm Pile-up 0, restricted to |eta| < 1. 4 εb= 0. 6 for udsg mistagrate of ~0. 55% εb= 0. 6 for udsg mistagrate of ~0. 6% Pile-up 100, restricted to |eta| < 1. 4 John Ellison & Ferdinando Giordano εb= 0. 6 for udsg mistagrate of ~10. 0% Pixel Upgrade Phase 1 Meeting, 27 April 2010 εb= 0. 6 for udsg mistagrate of ~4. 0% 14 H. W. K. Cheung (FNAL)

From Feb 11 Status Report: Tracking Efficiency and Fakerate Studies Full Sim using CMSSW_2_2_6 with iterative tracking and 3/4 seeding Four 100 Ge. V/c muons Using the version of FPix geometry known to have problems in negative eta (shall move to new geometry soon) Increasing the minimum number of hits per tracks from 3 may reduce the fake rate at high pileup ( as seen in Fast Sim with non-iterative tracking) Carlo Civinini Pixel Upgrade Phase 1 Meeting, 27 April 2010 H. W. K. Cheung (FNAL) 15

From Feb 11 Status Report: Seeding Efficiency and Fakerate Studies Full Sim using CMSSW_2_2_13 Comparing standard geometry and standard triplets (3/3) to Phase 1 geometry with iterative tracking and 3/4 seeding 100 Ge. V/c di-muons Using new barrel geometry from Xingtau and quick approach for FPix Alessia Tricomi Pixel Upgrade Phase 1 Meeting, 27 April 2010 H. W. K. Cheung (FNAL) 16 16

From Feb 11 Status Report: Tracking and Seeding Efficiency Studies Full Sim using CMSSW_2_2_13 comparison between No Pile. Up and Lumi 2*10^34 for the Phase 1 geometry 100 Ge. V/c di-muons using new barrel geometry from Xingtau and quick approach for FPix Alessia Tricomi Pixel Upgrade Phase 1 Meeting, 27 April 2010 H. W. K. Cheung (FNAL) 17 17

Current Simulation Studies Meeting agenda Better to refer you to talks tomorrow: • Redoing many of the previous studies and more with new Phase 1 geometry and with CMSSW_3_3_6 Pixel Upgrade Phase 1 Meeting, 27 April 2010 H. W. K. Cheung (FNAL) 18

Future Plan and task List Plan of work for Phase 1 Technical Proposal (Produced by V. Chiochia and M. Mannelli with input from A. Tricomi and H. Cheung) 1. Increased 4 th Layer Radius. (Motivations: Improve seed efficiency and quality, improve Tracking and HLT performance and increase robustness, improve Tracking performance in core of high-p. T jets) 1. Define seeding strategy for 4 layers (pairs, triplets, quadruplets) in the framework of iterative tracking 2. Plot seeding efficiency and track parameter resolution (vs eta/p. T) at any iterative step and total 3. Study combinatorics: seed multiplicity, timing performance, cut optimization 4. Introduce luminosity-induced (readout) inefficiency in the first layer 5. TIB search window and combinatorics in TIB 1/2/3 6. Tracking performance in core of high-p. T jets Pixel Upgrade Phase 1 Meeting, 27 April 2010 H. W. K. Cheung (FNAL) 19

Future Plan and task List 2. Reduced innermost radius and new pixel cell size. (Motivations: Improve/extend b-jet tagging capabilities, secondary vertices, and tau reconstruction) 1. Compare with current 3 -layer system 1. 2. 3. 4. Track impact parameter resolution vs momentum Primary and secondary vertex resolution, 3 D flight distance resolution b-tagging efficiency/purity Evaluate ROC/TBM data rates 2. Modify readout pitch in first layer (e. g. 75 100 m 2) and compare to 4 -layer system with standard pitch (100 150 m 2): 1. 2. 3. 4. 5. 6. Pixel position resolution vs. impact parameter Track impact parameter resolution vs. momentum Primary and secondary vertex resolution, 3 D flight distance resolution b-tagging efficiency/purity Seeding efficiency for dense environments (high p. T jets, tau’s) Evaluate data ROC/TBM data rates 3. Repeat some task in (2) for combination of smaller radius and smaller pixel cell size 1. Priority to data rates, b-tagging performance and seeding efficiency in dense jets (various Jet p. T) Pixel Upgrade Phase 1 Meeting, 27 April 2010 H. W. K. Cheung (FNAL) 20

Future Plan and task List 3. Reduced material budget. (Motivations: Improve Tracking efficiency and quality, reduce nuclear interactions, electron bremsstrahlung and photon conversions) 1. Compare with current 3 -layer system 1. Tracking efficiency and fake rate (e. g. efficiency for pion reconstructions 2. Number of photon conversions and nuclear interactions vs radius 3. Benchmarks of electron reconstruction n We are now starting to focus on this task list: Effort already on-going in some of the tasks but need more focus on the end product; also Phase 1 geometry only recently done and ready in CMSSW_3_3_6 1. 2. 3. Signing-up and assigning people to tasks (please email Alessia and Harry), as well as putting down people already working on various studies Now milestones for end of June and end of summer for (first) results Technical/software tasks still needed w w w Hit position and error for upgrade geometry/smaller pixel size Geometry variation for 4 th layer with increased radius MC Production: Need pileup samples for each geometry variation (Full sim) Pixel Upgrade Phase 1 Meeting, 27 April 2010 H. W. K. Cheung (FNAL) 21

Summary Lots of recent effort and progress on technical tasks for Phase 1 geometry ready for Full (and Fast) simulation in CMSSW_3_3_6 • Few technical tasks to improve simulations (especially Fast simulation) Direct/redirect all efforts to Phase 1 Simulation Studies Had already directed/redirected effort to Phase 1 technical tasks, now focusing on simulation studies • We have a task-oriented simulation studies list to focus our efforts • Need to get more people working on the various simulation studies (especially with the tight deadlines/milestones that have been set) We still need some effort on technical tasks: • Pixel hit position and resolution routine is crucial for accurate simulation/comparisons • Need geometry variation with increased radius for 4 th layer • Need effort on MC Production (Full Simulation) • Merge/Coordinate effort with tasks from upgrade physics studies from S. Dasu • Port to CMSSW_3_5_X/CMSSW_3_6_X so we have a SL 5 release – Work only with CMSSW_3_3_6 for the summer studies/results Pixel Upgrade Phase 1 Meeting, 27 April 2010 H. W. K. Cheung (FNAL) 22