SIDIS: MC 2006

mDSTs

• The mDSTs are produced at Lyon, in sub-directories of the /sps/compass disk.
  Some of the files have been moved to the tape system /hpss/in2p3.fr. The corresponding production directories are then kept, but housing only documentation.
• The mDST files proper are stored w/ a variety of other files, in particular the logs of the execution of the production jobs.

• Several sets are available, bearing the generic name m2006.FL.<distinctive_tag>:
  • Plain: /sps/compass/02/yann/m2006.FL.r507[.xxx].
    (moved to /hpss/in2p3.fr/home/y/ybedfer/mc/m2006.FL.r507).
  • Default: /sps/compass/02/yann/m2006.FL.d507[.xxx] w/ a "d" for default JETSET tuning.
    (moved to /hpss/in2p3.fr/home/y/ybedfer/mc/m2006.FL.d507).
  • RADGEN:
       There has been several successive productions of RADGEN (LEPTO+RADGEN, to be precise) files, over the course of the successive versions of the LEPTO+RADGEN produced by Andrea.
       These versions explore the parameter space of the RC business (i.e. form factors and structure functions describing elastic and inelastic lepton-Nucleon scattering). But they differ also in other aspects, as Andrea progressed in his understanding of the probelm, corrected a few things here and there, went back to the original source code in some instances, etc...etc... These source code differences should have a minor impact. But this has not yet been demonstrated...
       Versions (i) and (ii) correspond to the same parameter set. But their source codes differ. (I set out to produce (ii) when I discovered than I have lost the ability to duplicate what I had done for (i) and therefore could no longer extend it in a fully consistent way...)
    1. /sps/compass/02/yann/m2006.FL.R507[.xxx].bak = RADGEN files.
             They have been copied to tape, see: /hpss/in2p3.fr/home/y/ybedfer/mc/m2006.FL.R507.
    2. /sps/compass/02/yann/m2006.FL.+507[.xxx] = New RADGEN files.
       (moved to /hpss/in2p3.fr/home/y/ybedfer/mc/m2006.FL.+507).
    3. /sps/compass/02/yann/m2006.FL.R507[.xxx], w/ a capital "R" but w/o the .bak extension, are produced w/ a, tentative, new version of RADGEN: not recommended!
       (moved to /hpss/in2p3.fr/home/y/ybedfer/mc/m2006.FL.R507).
    4. /hpss/in2p3.fr/home/y/ybedfer/mc/m2006.FL.T507[.xxx], w/ a capital "T" for radgen w/ TERAD inputs.
       (moved to /hpss/in2p3.fr/home/y/ybedfer/mc/m2006.FL.T507).
       
       
  • /sps/compass/02/yann/m2006.FL.r508[.xxx].bak = old TERAD files.

LEPTO

• Executable (at Lyon) in $THRONG_DIR/sl6/generators.
• Options:
  • default:   ~ybedfer/geant.utils/lepto.MSTW08.psON.opt,v1.1.
  • d507:      ~ybedfer/geant.utils/lepto.MSTW08.psON.default.opt,v1.1.
  • R507:      ~ybedfer/geant.utils/leptorad++.MSTW08.psON.opt,v1.1.
  • T507:      ~ybedfer/geant.utils/leptorad++.MSTW08.psON.TERAD.opt,v1.1.

COMGeant

• SVN r507    (or r508, which differs only by its truncated target cells)
• Options = ~ybedfer/geant.utils/geant.FLUKA.2006.opt,v1.1.

Target shifts

• In order to reproduce the real life position, see Korzenev, am_070802.
• Mainly in the vertical dimension, as per r506.
• Nota bene:In order to implement these shifts into the target cuts at the PHAST level, one cannot rely on PaAlgo::GetTargetLocation and PaAlgo::CrossCells (which set the target parameters based on built-in values),
  => one has to devise one's own methods, see e.g. ~ybedfer/phast.utils/user/DISTarget.

Hadron reinteraction

• FLUKA is used, instead of GHEISHA.
• FLUKA is believed to better reproduce real data as a function of p_T, see All p_T Q²>1 2014 release note.
• It's brought into play by option HADR 4, thanks to r504. Note that option MUNU is then disabled.
• This, based on previous work by Sebastian Uhl reported here.

Coral

• prod-2006-12-22-slc6, r13927.
• Justification: see coral versions of 2006 mass prod.s.
• Options = ~ybedfer/geant.utils/trafdic.mc.06W41.opt.
  • It's v1.1 until /sps/compass/02/yann/m2006.FL.r507.3e, included.
  • And then v1.2 w/ MC generator info output to mDST's.
    So is it for RADGEN and LEPTO with default JETSET infra.

RADGEN

• Temporarily, the distribution is stored at Lyon on sps/compass/02/yann/extgen_cpp.
  It's under the git version control system, see: https://gitlab.cern.ch/compass_extgen/extgen_cpp.
• Else, it's $THRONG_DIR/sl6/generators/sl6/generators/mlepto_rad.
• Two successive rounds:
  1. Standard, options = ~ybedfer/geant.utils/leptorad.MSTW08.psON.opt,v1.1.
  2. W/ TERAD inputs, options = ~ybedfer/geant.utils/leptorad++.MSTW08.psON.opt,v1.1.
• Coral options: v1.2, w/ MC generator info.

LEPTO with default JETSET

• Options = ~ybedfer/geant.utils/lepto.MSTW08.psON.default.opt,v1.1.
• Coral options: v1.2, w/ MC generator info

Standard LEPTO (w/ high-pT tuning) with truncated target cells.

• So far only one set of 512 files of 8000 events produces.
• It's a modified r508. The changes w.r.t. r507, which are in fact committed as r509, are the following:

622a623
> C Filling of the target: see  infra.
658,661c659,670
<   PARVOL80 'UPST'  252  'TA50'     0.    0.   0.   0  'TUBE'  3     0.0   1.5    15.    LiD
<   PARVOL81 'MDDL'  252  'TA51'     0.    0.   0.   0  'TUBE'  3     0.0   1.5    30.    LiD
<   PARVOL82 'DNST'  252  'TA52'     0.    0.   0.   0  'TUBE'  3     0.0   1.5    15.    LiD 
< C
---
> C (I(Y.B. dont't understand this last comment about "Packing Factor")
> C - In the present "alt2006Target" branch of SVN, we shorten the 3 cells (2006)
> C  by resp. (4,8,2 cm) for (u,c,d), see: "am_070802/korzenev_070801_2.pdf" on
> C  "wwwcompass.cern.ch/compass/software/analysis/transparencies/2007/"
>   PARVOL80 'UPST'  252  'TA50'     0.    0.    2.   0  'TUBE'  3    0.0    1.5    13.    LiD
>   PARVOL81 'MDDL'  252  'TA51'     0.    0.    4.   0  'TUBE'  3    0.0    1.5    26.    LiD
>   PARVOL82 'DNST'  252  'TA52'     0.    0.    1.   0  'TUBE'  3    0.0    1.5    14.    LiD 
> C Add 3 disks of He, filling the space left by the shortened LiD 
>   PARVOL84 'HEUP'  263  'TA50'     0.    0.  -13.   0  'TUBE'  3    0.0    1.5     2.    LHe
>   PARVOL85 'HEMD'  263  'TA51'     0.    0.  -26.   0  'TUBE'  3    0.0    1.5     4.    LHe
>   PARVOL86 'HEDN'  263  'TA52'     0.    0.  -14.   0  'TUBE'  3    0.0    1.5     1.    LHe 
> CC

-- YannBedfer - 2014-11-06