ALICE Off-line Computing: Status and Plans P.Hristov for the ALICE Collaboration 14/09/2007 NEC’2007, Varna
Jan 14, 2016
ALICE Off-line Computing: Status and Plans
P.Hristov for the ALICE Collaboration
14/09/2007
NEC’2007, Varna
NEC'2007, 14/09/2007 P. Hristov 2/21
History
NEC’2001 AliRoot for simulation NEC’2003 Reconstruction with AliRoot NEC’2005 AliRoot for analysis NEC’2007 Still no LHC data => Status and
plans of the ALICE offline software
NEC'2007, 14/09/2007 P. Hristov 3/21
Outline
ALICE Experiment Simulation Reconstruction Analysis First physics Conclusions
NEC'2007, 14/09/2007 P. Hristov 4/21
ALICE detectors
L3 magnet B<0.5 T HMPID TRD Dipole magnet
MUON spectrometer
Absorber
TPCPHOS
ITS
PMD
TOF
NEC'2007, 14/09/2007 P. Hristov 5/21
Offline framework AliRoot in development since 1998
Directly based on ROOT Used since the detector TDR’s for all ALICE studies
Few packages to install: ROOT, AliRoot, G3/G4/FLUKA, AliEn client Ported on most common architectures/OS
Linux i686, ia64 and x86_64, Mac OS X, Digital True64, Solaris x86,… Distributed development
Over 50 developers and a single CVS repository 2/3 of the code developed outside CERN
Integration with DAQ (data recorder) and HLT (same code-base) Abstract interfaces and “restricted” subset of C++ used for maximum
portability
NEC'2007, 14/09/2007 P. Hristov 6/21
Software management Regular release schedule
Major release every two months, minor release (tag) every month Nightly produced UML diagrams, code listing, coding rule violations,
build and tests , single repository with all the code No version management software (we have very few packages!)
Advanced code tools under development (with IRST/Trento) “Smell” detection Aspect oriented programming tools Automated genetic testing
Documentation for AliRoot and AliEn produced Intensive training ongoing
NEC'2007, 14/09/2007 P. Hristov 7/21
The Simulation
User Code
VMC
Geometrical Modeller
G3 G3 transport
G4 transportG4
FLUKA transportFLUKA
Reconstruction
Visualisation
Generators
NEC'2007, 14/09/2007
P. Hristov 8
Event Generation:Kinematics
Particle Transport:Hits (energy
deposition at givenPoint, MC label)
Detector Response:Summable Digits
(low ADC threshold,no noise, MC label)
Detector Response:Digits
(noise + normal Threshold, MC label)
Raw data formats:DDL files, DATE file,
“rootified” raw
Event Merging (MC + MC): Summable Digits Signal+Summable Digits BackgroundNeeded to reduce the simulation time
Event embedding (MC + Raw): Summable Digits Signal+Raw converted to SDigitsStudies of reconstruction efficiency
Event mixing: tracks or clusters fromone event are combined with tracksor clusters from another (but “similar”)event.
SIMULATION
Reconstruction
NEC'2007, 14/09/2007 P. Hristov 9/21
Visualisation
QuickTime™ and aTIFF (Uncompressed) decompressor
are needed to see this picture.
QuickTime™ and aTIFF (Uncompressed) decompressor
are needed to see this picture.
QuickTime™ and aTIFF (Uncompressed) decompressor
are needed to see this picture.
QuickTime™ and aTIFF (Uncompressed) decompressor
are needed to see this picture.
NEC'2007, 14/09/2007 P. Hristov 10/21
Plans - AliRoot Framework
Improvements in the CPU and the memory consumption Ongoing redesign of the ESDs classes Finalize common AOD classes Data size reduction: ESD, AOD Optimize access to OCDB & metadata selection Simplification of the IO management RAW data: take into account the latest modifications in the hardware
Simulation Urgent – document & implement ALICE geometry as built Material in tracking region & fwd direction (services, cables) Getting TFluka & G4 production ready
Visualization Preparation for event scanning and visual quality assurance
NEC'2007, 14/09/2007
P. Hristov 11
Local DetectorReconstruction:
Clusterization, vertex
Seeding:2 clusters in TPC+ primary vertex
Kalman Filter:Forward propagation
TPCITS
Kalman Filter:Backward propagationITSTPCTRDTOF
Kalman Filter:Refit inward
TRDTPCITS
RECONSTRUCTION
BARREL
TRACKING
TPC seeding,Forward tracking
ITS forward tracking,Combinatorial Kalman
filter
ITS forward tracking
TPC V0 and kinkfinder
ITS V0 finder
ITS tracking
TPC tracking
TRD tracking, seeding
Update ofV0s and kinks
TRD tracking, PID
TPC tracking, PID
ITS tracking, PID
Update ofV0s and kinks
Update ofV0s
TOF PID
Propagation to PHOS,EM,HMPID
NEC'2007, 14/09/2007 P. Hristov 12/21
Tracking efficiency
TPC
all detectors
For realistic particle densitiesdN/dy = 2000 – 4000
combined efficiency well above 90%and fake track probability below 5%
Challenge in high-particle density environment
Efficiency normalized to number of
“reconstructable” track
NEC'2007, 14/09/2007 P. Hristov 13/21
Combined efficiencyEfficiency normalized to number of
generated particles at primary vertex
within the central acceptance |η|<0.9 protons – absorption kaons and pions – decays
central Pb – Pb TPC
At lower pt : TRD inclusion dropsthe efficiency due to unavoidable material between TRD sectors
pp
NEC'2007, 14/09/2007 P. Hristov 14/21
Plans – Reconstruction Alignment-aware reconstruction
Survey data usage Development and tests of the alignment algorithms
Calibration-aware reconstruction Ongoing review detector by detector Development of detector algorithms
Reconstruction parameters in separate classes Factorization of reconstruction features
Magnetic Field Ongoing analysis of the measurements, close to completion
PID: consistent treatment of the mismatch between detectors
NEC'2007, 14/09/2007 P. Hristov 15/21
AlignmentSimulation
Ideal Geometry
Misalignment
Reconstruction
Raw data
File from survey
Ideal Geometry
Alignment procedure
NEC'2007, 14/09/2007 P. Hristov 16/21
OCDBGrid File Catalog
DIMtriggerECS
RunLogbook
DAQ
FXS DB
FXS
DCSarch. DB
DCS
FXS DB
FXS
HLTFXS DB
FXS
SHUTTLE
TR
D
MP
H
SP
D
TP
C
...Calibration: The Shuttle
No alternative system to extract data (especially online calibration results) between data-taking and first reconstruction pass!
NEC'2007, 14/09/2007 P. Hristov 17/21
High-level dataflow
CAF
WN{se’s}guid lfn
{se’s}guid lfn
{se’s}guid lfn
{se’s}guid lfn
guid {se’s}lfn
ALICE File Catalogue
Publishagent xrootd
Castorcache
CASTOR
LDC
DAQNetwork
GDC
Conditionfiles
Data files
Data fileFTS
SRM
SRM
T1’s
OfflineOnline
HLT
DDL
240TB
DCS Shuttle
Publish in AliEn
Monit.Calib.
DAQFXSDAQ
Logbook DB
Condition files
HLTFXSDCS
FXS
Run info
DCSDB
NEC'2007, 14/09/2007 P. Hristov 18/21
p
ALICE Analysis Basic Concepts Analysis Models
Prompt data processing (calib, align, reco, analysis) @CERN with PROOF
Batch Analysis using GRID infrastructure Local analysis Interactive analysis PROOF+GRID
User Interface Access GRID via AliEn or ROOT UIs
PROOF/ROOT Enabling technology for CAF GRID API class TAliEn
Analysis Object Data contain only data needed for a particular analysis
Analysis à la PAW ROOT + at most a small library
Work on the distributed infrastructure has been done by the ARDA project
NEC'2007, 14/09/2007 P. Hristov 19/21
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
CAFlfn guid {se’s}
lfn guid {se’s}
lfn guid {se’s}
lfn guid {se’s}
lfn guid {se’s}
xrootdCASTOR
WNPROOFXROOTD
WNPROOFXROOTD
WNPROOFXROOTD
WNPROOFXROOTD
WNPROOFXROOTD
WNPROOFXROOTD
PROOF master
QuickTime™ and aTIFF (Uncompressed) decompressorare needed to see this picture.
The whole CAF becomes a xrootd cluster
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
For prompt and pilot analysis, calibration and alignment, fast simulation and reconstruction
Working with the PROOF team on optimisation S/W versions handling now in PROOF Disk quotas are already partly implemented and
under test Data distribution at the moment “by hand" Fairshare algorithm under development
NEC'2007, 14/09/2007 P. Hristov 20/21
First Physics Studies:Measuring the Pseudorapidity Density of Primary Charged Particles using the TPC (ALICE-INT-2007-005)
Event and track quality selection. E.g. Events with reconstructed
vertex Require track origin in
proximity to the primary vertex Corrections
Track-to-particle (track level):f(, z vtx, pT) Geometrical acceptance, reconstruction efficiency, decay, feed-down
Vertex reconstruction efficiency (event level) f(z vtx, multiplicity)
Trigger efficiency (event level) Corrects to INEL, NSD f(z vtx, multiplicity)
Studies of systematics Track selection Influence of secondaries Particle composition Pt cutoff Relative frequency of processes Beam-gas and pileup events Misalignment Bias from the vertex reconstruction
PDC06 data:400K events (corrections)100K events (analysis)Different event sample for corrections and analysis!
NEC'2007, 14/09/2007 P. Hristov 21/21
Conclusions The ALICE offline software is making steady
progress towards readiness Well prepared for real data and first physics studies Optimization of the code performance (CPU, memory,
file size) Last chance to add missing functionality Extensive tests during the detector commissioning
The area between online and offline is the most demanding at the moment Contingency is small but positive