GEANT 4 - MOU

1
LHCb Computing(DAQ, ECS, Software) RD and
Milestones
Meeting with LHCb Referees July 8th, 1998
2
Overall Schedule for DAQ and ECS
3
Goals of RD Phase

• Devise an architecture for the DAQ system and a
  specification for all dataflow and control
  elements
• Acquire knowledge of, and experience with, new
  technologies
• Assemble small scale hardware prototype of DAQ
  system (String Test) running at full speed
• Finally take an educated decision on the
  technologies to use for the implementation of the
  final system

4
DAQ Activities

• Architecture and protocol design
• Readout Unit Implementation Study
• Study functionality, interfacing
• Design and prototype, performance
• Event Building Project
• Devise strategy L1,L2/3
• Study technologies e.g. Myrinet
• Simulation models, demonstrators
• Timing and Fast Control
• Readout Supervisor
• FEM Implementation Study
• Event Filter Farm Study (LCB Project)
• Study capabilities of Mass storage (ALICE/IT)

5
Experiment Control System (ECS)

• Joint Project to devise a common controls kernel
  for all LHC experiments and all aspects of
  control
• Selected Sub-Projects .
• Joint URD for Alice/LHCb (finished)
• Hardware interface URD (need SD input)
• Architecture design (ongoing)
• Technology survey (finished)
• CANbus evaluation (ongoing)
• OPC evaluation (starting after summer)

6
Schedule for DAQ/ECS
7
Where are we gOOing?
June 1998
June 2000
OO Frameworks
SICB
OO toolkits
FORTRAN Toolkits
OO event geometry models
ZEBRA banks
ZEBRA and ASCII files
OO database (Objectivity/DB)
8
Schedule for Computing
9
Plans for SICB

• SICB will be discarded when new software with
  same or superior functionality will be available
• Current planning July 2000
• Until then
• Production simulation will continue with SICB
• Store event generator record to be available for
  analysis
• Interface to other event generators
• Studies of alternative detector layout options
• Enhanced detector response simulation

10
Working Prototype

• By end 1999, produce a new working prototype of
  LHCb software incorporating
• model of the LHCb detector geometry (-gt technical
  forum)
• database containing 1,000,000 simulated events
  (100 GB)
• framework for processing events (Components,
  integration technology)
• detector response simulation toolkit (GEANT4)
• pattern recognition algorithms for tracking,
  RICH,..
• toolkit of analysis algorithms
• Compare toolkits for interactive data analysis
  (PAW, ROOT, LHC..)

11
Event Model and Analysis Framework

• Goals - short term
• Provide access to SICB data from C
  (SICB)(mid-July)
• Collect requirements for transient event object
  model (summer)
• Writing processed data back to ZEBRA files is NOT
  a priority

Analysis application
Exists using SICB event loop
Event loop
LHCb OO transient event
Interface
Start with Event header class
SICB C classes
Iain Last package
Exists, two weeks to make sure it works correctly
Zebra file
12
Software Development

• Documentation templates for user requirements,
  project plans, costings..
• Choose a method and tool for designing LHCb
  software
• Evaluate PC platforms (NT, Linux)
• Support for coding and testing (compilers,
  debuggers)
• Develop code management and software release
  scheme
• Develop automatic code documentation scheme

Work with LHC SDE project (led by IT/ IPT
group) Need to develop a core of experienced OO
programmers First SDE definition by September
98 Need a mature process and SDE by end 99
13
Computing Model

• Compute facilities - PC farms running NT or Linux
• Marseilles, Liverpool, Rio, with other LHC
  experiments and IT/PDP group
• Data Management and Computing Using Distributed
  Architectures
• with other LHC experiments in a proposed LCB
  project
• outside institutes CERN/IT (LHCb/Oxford,)
• determine which classes of models for distributed
  data analysis are feasible, taking into account
  network capabilities and data handling resources
  likely to become available in the collaboration
  sites
• identify and specify the main parameters and
  build tools for making simulations for comparison
  of alternative strategies.
• Make test implementations of elements of
  computing model