LHCb Computing Status Report DAQ , ECS , Software, Facilities

1
LHCb Computing Status Report DAQ , ECS ,
Software, Facilities

• John Harvey
• CERN / EP
• Meeting with LHCC Referees
• 27 - Nov 2000

2
LHCb Trigger/DAQ/ECS Architecture
3
LHCb TFC system

• Readout and Throttle Switches
• Programmable for Partitioning
• Design reviewed in October 00
• Prototypes in February 01
• Readout Supervisor
• Functional Specification done
• Design in progress
• Review scheduled for April 01
• Prototype scheduled for Oct 01
• TFC system test
• Acquire components
• Test L1 broadcast via channel B
• Start June 01

4
LHCb Readout Unit

• New version of RU design (v2)
• less chips, fewer layers, lower cost
• First prototype in first week of December 00
• Programming of FPGA code for implementing readout
  protocols underway
• Working and tested modules expected by end Jan
  01
• Integration tests will start Mar 01

5
LHCb Event Builder

• Studied Myrinet (buffers needed)
• Now focusing on Gbit Ethernet
• Test setup between 2 PCs
• Use gt95 nominal bandwidth for frames gt512 bytes
  (512 -gt230 kHz)
• Can send out frames at frequencies up to 1.4 MHz
  for 64 byte frames
• Implement event building in NIC
• Frequency of 100 kHz demonstrated
• EB at Gbit speeds for gt 1kB frames demonstrated
• Tested 1 on 1 event building over a switch in CMS
  test bed
• Fixed protocol with RU
• Presented results at DAQ2000
• Now studying flow control in switch and making a
  full scale test on CMS test stand (16 on 16)
• More detailed simulation of a full scale GbE
  readout network to be done

CERN Network
6
ECS interface to Electronic Modules

• Select a reduced number of solutions
• Support (HW and SW) for the integration of the
  selected solutions
• Ethernet to credit card PC for use in counting
  room
• Test board being developed
• Study i/f of CC-PC to Parallel bus, I2C, JTAG
• Test functionality (RESET) and measure noise
• Prototype expected in January 01
• Results end of March 01
• N.B. two other solutions considered for use in
  high level radiation areas
• SPAC Long Distance I2C/JTAG
• CMS tracker CCU

7
ECS Control System Prototype

• Aim is to distribute with the SCADA license a
  framework where users can easily implement
  sub-systems.
• First prototype comprises
• Configuration rules and conventions (naming,
  colors, etc)
• Tools for Device Integration
• Hierarchical Control Partitioning
• Based on Finite State Machines and SCADA
• Automatic UI Generation
• Based on SCADA
• Plans are to use prototype in the LHCb test beam

8
Online team members
Leaving In 2000

• DAQ
• Beat Jost project leader CERN staff
• Jean-Pierre Dufey Readout Network CERN staff
• Marianna Zuin Readout Network technical
  student
• Richard Jacobsson TFC CERN staff
• Niko Neufeld Readout Network CERN Fellow
• EP/ED group Readout Unit CERN staff
• Zbigniew Guzik engineer Warsaw
• ECS
• Clara Gaspar project leader CERN Staff
• Wolfgang Tejessy JCOP CERN Staff
• Richard Beneyton SCADA in test beam Cooperant
• Sascha Schmeling SCADA test beam CERN Fellow

Arriving in 2000
9
Software Framework - GAUDI

• GAUDI v6 released on Nov 10th
• Enhanced features e.g. event tag collections, XML
  browser, detector geometry and event display
• 110,000 lines of code, 460 classes
• Good collaboration with ATLAS
• New services by ATLAS auditors, histograms in
  ROOT, HEPMC
• Moving to experiment-independent repository
• The most urgent tasks for next release are
• Event model with subdetector groups
• Detector description with subdetector groups
• Conditions database with CERN/IT
• Consolidation and enhancements (code
  documentation Doxygen)
• Further contributions expected from ATLAS
• Scripting language for interactive work
• HARP, GLAST and OPERA also users of GAUDI

10
Software Applications

• GAUDI-based event reconstruction (BRUNEL)
• BRUNEL v1 r5 released this month
• Physics functionality entirely based on wrapped
  FORTRAN code
• First public release of C track fit integrated
  tested
• Pile-up implemented and spill-over being
  implemented
• Available for production tests
• Migration of detector software to C
• Progress in all areas - digitisation, geometry
  description,
• Tracking - digitisation and pattern recognition
  almost ready for public release
• e.g. VELO, CAL complete event model and
  geometry description
• Current activities reflect TDR schedule
• VELO, MUON (and Tracking) now on hold until after
  TDRs produced
• RICH and CAL new software getting higher
  priority

11
Software Applications GEANT4

• Developing interface to G4 for GAUDI applications
  (GiGa)
• Isolates G4 code from GAUDI
• Way to input detector geometry and kinematics to
  G4
• Handles passing of commands to G4 and retrieval
  of events from G4
• More generally Geant 4 physics being tested
• Now - by BaBar, ATLAS, ALICE, Space applications
• Some disagreements with data and G3 seen and
  being studied
• Plans in LHCb
• Calorimeter simulation of shower production in
  prototype and comparison with existing
  simulations (G3) and data from testbeam
• RICH studying production and detection of
  Cherenkov photons in RICH1 using TDR geometry
  compare results
• Integration of these developments in GAUDI using
  GiGa
• Measure performance

12
LHCb CORE Software Team
Left In 2000

• Pere Mato project leader CERN staff
• Florence Ranjard code librarian CERN staff
• Marco Cattaneo BRUNEL CERN staff
• Agnieszka Jacholowska - SICb Orsay
• Markus Frank GAUDI CERN staff
• Pavel Binko GAUDI CERN staff
• Rado Chytracek GAUDI doctoral student
• Gonzalo Gracia GEANT4 CERN fellow
• Stefan Probst - GAUDI technical student
• Gloria Corti GAUDI CERN fellow
• Sebastien Ponce - GAUDI doctoral student
• Ivan Belyaev (0.5) GAUDI/GEANT4 ITEP
• Guy Barrand (0.5) event display ORSAY

Arriving in 2000
13
Computing Facilities

• Estimates of computing requirements updated and
  submitted to Hoffmann LHC Computing Review
• NT farms being decommissioned at CERN and at RAL
• Migrating production tools to Linux now
• Starting production of 2M B-inclusive events at
  Liverpool
• Farm of 15 PCs for LHCb use at Bologna early 2001
• Long term planning in INFN going
  on(location,sharing etc)
• Farm of 10 PCs to be set up early 2001 in NIKHEF
  for LHCb use
• Developing overall Nikhef strategy for LHC
  computing /GRIDs
• Grid Computing in LHCb
• Participation in EU Datagrid project starts Jan
  2001 (3 years)
• Deploy grid middleware (Globus) and develop
  production application
• Started mini-project between Liverpool/RAL/CERN
  for testing remote production of simulation data
  and transfer between sites

14
LHCb Computing Infrastructure Team

• Frank Harris coordination Oxford
• Eric van Herwijnen MC production/Grid CERN
  staff
• Joel Closier system support/bookkeeping CERN
  staff