Q: What gcc version / ROOT version / Linux version I need to run PHAST ?
A: Here is not complete list of platforms where PHAST had been tested and used:
- gcc 2.95.2 ROOT v3.00.06 RedHat 6.1 (LXPLUS) - gcc 2.95.2 ROOT v3.03.09 RedHat 7.3 (LXPLUS) - gcc 2.95.2 ROOT v3.05.05 RedHat 7.3 (LXPLUS) - gcc 2.96 ROOT v3.03.09 RedHat 7.2 - gcc 2.96 ROOT v3.05.02 RedHat 7.2 - gcc 3.2 ROOT v3.05.02 RedHat 8.0 - gcc 3.2.2 ROOT v3.10.01 RedHat 9.0 - gcc 3.2.2 ROOT v4.00.08 RedHat 9.0 - gcc 3.2.3 ROOT v4.01.02 SLC3 (Scientific Linux at CERN) - gcc 3.2.3 ROOT v4.03.01 SLC3 (Scientific Linux at CERN) - gcc 3.2.3 ROOT v4.03.04 RedHat 9.0 - gcc 3.2.3 ROOT v4.03.04 SLC3 (Scientific Linux at CERN) - gcc 3.2.3 ROOT v5.02.00 ES3A - gcc 3.2.3 ROOT v5.02.00 ES3A - gcc 3.4.5 ROOT v5.10.00c SLC4 (Scientific Linux at CERN) - gcc 3.4.6 ROOT v5.14.00b SLC4 (32-bit and 64-bit mode) - gcc 4.1.2 ROOT v5.24.00 SLC5 (32-bit and 64-bit mode) - gcc 4.3.2 ROOT v5.28.00b SLC5 (32-bit and 64-bit mode)
Q: How to compile Phast on platform different from default one
(lxplus.cern.ch)
A: Have a look on set of Makefile.* in Phast directory, find
what corresponds to your system configuration, change symbolic
link Makefile:
to corresponding Makefile.xxxx and make.
Q: Is it possible to install PHAST on system without Fortran compiler, without CERNLIB and/or
without RFIO
A: Yes. Use following options of make:
make NO_FORTRAN=1 NO_RFIO=1
Q: On SLC4 machine make ends up with following error:
/usr/bin/ld: cannot find -lRFIO
What shell I do?
A: It means that ROOT is compiled without RFIO support (i.e. there is no $ROOTSYS/lib/libRFIO.so)
On SLC4 platform SHIFT library (RFIO) is included in the system (/usr/lib/libshift.*), so ROOT with RFIO support is not needed anymore.
Compile with
make NO_RFIO=1
Q: I can't link Phast in 64 bit mode
A: Phast still have functions which use CERNLIB
(e.g. PaTPar::FindCDA()) but CERNLIB is provided as 32 bit library only. So, try
make NO_FORTRAN=1
to disable use of Fortran code and CERNLIB. If your code calls one of disabled
functions you will get an error message.
Q: How to compile PHAST with optimization ?
A: Default compilation option (starting from v. 7.015) is "-O2".
To compile with "debug information" :
make OPT=-g
Other optimization levels could also be specified. (e.g. make OPT=-O1)
(Remark: do not use level "3", as some of ROOT headers are not compatible with this optimization level)
$ make clean $ makeIf it didn't help, try to take new PHAST version. There is non-zero probability, that problem you discovered had been found and fixed already.
Q: Which units are used in PHAST?
A: [cm], [GeV], [ns], [Tesla]
Q: What is the coordinate system?
A: Right-hand system: Z - along the beam, Y - upward, X - from Saleve to Jura
Origin is at the Polarized Target center.
But it is "nominal" target center. E.g. "Saclay" magnet was moved 35 cm upward.
For setups of COMPASS hardon program, origin is the same as for muon program.
Q: I can't read new mDST with my PHAST which works fine with old mDSTs.
A: This can happen if new mDST had been produced with ROOT version newer than version
you use for PHAST compilation.
Unfortunately, ROOT versions are not always "forward compatible".
So, switch to newer ROOT version (i.e. redefine your $ROOTSYS and $LD_LIBRARY_PATH) and do
make clean ; make
Q: What is called primary vertex in PHAST?
A: It's a vertex which contains beam track as "incoming" track. Nothing more. Event may contain
more then 1 primary vertex.
Q: I see that for the beam track functions ZLast() and ZFirst() (last and first measured point)
return the same coordinate. Why?
A: ZFirst() take coordinate from the first saved track parameter (vTPar(0)) while Zlast() returns Z position of detector,
corresponding to last bit in track's hitmap.
For beam tracks order of track parameters was reversed to have first one closest to target, while track's hitmap stay unchanged.
If it is too confusing, use Zmin()/Zmax() function which returns Z position of most upstream/downstream detector
where this track has hits.
Q: Which particles are tagged as mu prime (i.e. PaParticle::IsMuPrim() function returns "true") ?
A: Particle which is NOT the beam and tagged as "SPECIAL" in CORAL (ask muon ID expert(s) for more details).
Event may contain few mu' particles. Even one vertex may contain few particles tagged as mu'.
Q: Why PaParticle::PID() always returns 5 or -1 ?
A: Currently, only muon identification is rather unambiguous, so PID could be 5, 6 or -1 (unknown)
RICH information is also available on PaTrack, but it's interpretation is left to the end user.
Q: Which particles are identified as muon?
A: Particles already tagged as mu prime in CORAL OR if fraction of radiation lengths
traversed by corresponding track X/X0 > 30.
Q: How to pass counters from UserEventN() to UserJobEndN() ?
A: Recommended way: book histogram for counters (1 bin - 1 counter) and fill it in UserEvent().
ROOT histograms could be "seen" everywhere in the code (see UserJobEnd0.cc as example).
How to access bin contents, consult ROOT manual.
Q: What is the meaning of RICH information one can get by RichInf() function of PaTrack
(NRichInf() elements in total)?
A: This information is just a copy of an array stored in CORAL's CsTrack.
It's better to ask RICH experts about exact meaning of every value.
For current version of RICH package, content of RICH block is following:
//- partProbs[ 0 ] background Like
//- partProbs[ 1 ] pion Like
//- partProbs[ 2 ] kaon Like
//- partProbs[ 3 ] proton Like
//- partProbs[ 4 ] pion Like derivative
//- partProbs[ 5 ] kaon Like derivative
//- partProbs[ 6 ] proton Like derivative
//- partProbs[ 7 ] maximum likelihood angle
//- partProbs[ 8 ] reconstr. ring angle
//- partProbs[ 9 ] number of photons per part-photon or ring
//- partProbs[ 10 ] fitted ring angle
//- partProbs[ 11 ] ring chisquare
//- partProbs[ 12 ] pion chisquare
//- partProbs[ 13 ] kaon chisquare
//- partProbs[ 14 ] proton chisquare
//- ---------------------------------- added
//- partProbs[ 15 ] electron Like
//- partProbs[ 16 ] muon Like
//- partProbs[ 17 ] electron Like derivative
//- partProbs[ 18 ] muon Like derivative
//- partProbs[ 19 ] electron chisquare
//- partProbs[ 20 ] muon chisquare
(Paolo.Schiavon@cern.ch, 11.04.2005, "coral-weekly" mailing list)
Q: How to get in PHAST refractive index of RICH radiator for run I process ?
A: There are few methods to do this:
1) PaMetaDB::Ref().NminusOne(run) return refractive index for the run # "run" as it was found in calibrations DB at the date which one could see in the first line of ./src/RICH1_ref_index.db file. Normally this "DB snapshot" is updated with every PHAST release.
2) PaSetup::Ref().NminusOne() returns refractive index used at production time. (Unfortunately it was not stored in mDST of early 2002 data production. To recover this information please use special function PaMetaDB::Ref().NminusOneProd(run) instead).
Q: How to estimate beam flux, which corresponds to processed event sample?
A: 2 methods of the flux estimation had been implemented in 2 functions:
float Phast::IntegratedBeamFlux()
float Phast::TotalBeamFlux()
See the class documentation for more details
Q: I would like to look on the scattered muon at the place downstream the second
muon absorber, is there any track information behind SM2 available?
A: In "standard" miniDST from the production this information is not available.
Only trajectory parameters in first measured point of track are stored.
But if precision and speed is not very crucial, one may extrapolate available trajectory parameters,
to Z position you interesting in (see function PaTPar::Extrapolate()).
From the other hand, user may create his own mDST (see ./coral/README) with all avalable parameters
of trajectory. If storage of some hits will be requested, this mode is ON automatically.
Q: What is G3partMass[i] in UserEvent0.cc?
A: It's particles table
Q: I would like to look on the scattered muon at the place downstream the second
muon absorber, is there any track information behind SM2 available?
A: In "standard" miniDST from the production this information is not available.
Only trajectory parameters in first measured point of track are stored.
But if precision and speed is not very crucial, one may extrapolate available trajectory parameters,
to Z position you interesting in (see function PaTPar::Extrapolate()).
From the other hand, user may create his own mDST (see ./coral/README) with all available parameters
of trajectory. If storage of some hits will be requested, this mode is ON automatically.
Q: I work with "MegaDST" (i.e. miniDST with additional hits information).
Looking on hits from detectors with F1 chip readout sometimes one may observe more than
1 digit per wire and sometimes they are identical. Is it OK?
A: Yes, it's OK. There are 2 explanations:
Q: What is pileup MC vertex ?
A: 2 types of vertices are tagged as "pileup":
Q: What are strange numbers PaMCTrack::Pid() function returns?
A: It's particle ID in the notation similar to
GEANT particles table (the same about PaParticle::PID() function).
Q: Why MC vertex tagged as "primary" do not have "incoming" track
(iTrackOfOrigin() == -1)? And where is MC beam track?
A: PaMCvertex structure is not "symmetric" to PaVertex as it reflects the structure of vertexes, generated by COMGEANT. One feature of the simulated beam is that the beam is normal
"outgoing" track in the primary MC vertex with the only difference that it goes "upward" (i.e. Pz < 0).
To get index of the beam track in MC vertex, one may use PaMCvertex function iBeam().
Currently, it's aways the first in vertex.
Q: In my kinematics analysis I use 3-momentum of MC beam track (PaMCtrack::Mom3()).
As MC beam goes "upward", should I multiply by "-1" all
3 components of MC beam track momentum or only Pz ?
A: One has to "invert" all 3 components. But if you use PaMCtrack::ParInVtx() to access
MC beam track parameters, nothing has to be done as parametrization used in PaTPar do not "know"
direction by definition.
The StreamerInfo of class PaParticle read from file ... has the same version (=8) as the active class but a different checksum. You should update the version to ClassDef(PaParticle,9). Do not try to write objects with the current class definition, the files will not be readable.What it means?
Q: At reading of CASTOR files PHAST crashes sometimes and produces a long list of error
messages starting from :