MSDP Workshop PRACTICAL SESSION P. Maeght, P. Mein N. Mein, M. Lafon, N. Meunier T. Roudier MSDP Workshop, Tarbes, 18-20th January 2006.

MSDP WorkshopPRACTICAL SESSION

P. Maeght, P. MeinN. Mein, M. Lafon, N. MeunierT. Roudier

MSDP Workshop, Tarbes, 18-20th January 2006

Directories oceanide

/data3/DPSMWS/day/ : raw sequences /data4/groupX/ (one dir. / group, X=1,2)

auto0503/ : code for THEMIS observations until 2004

auto0601/ : code for THEMIS observations since 2005

dalsa/ : code for VTT docdpsm/ dpsm_form/

Results will be in auto0503/ or auto0601/

Directories gaia

/baie1/DPSMWS/day/ : raw sequences /baie1/groupX/ (one dir. / group, X=3,4)

auto0503/ : code for THEMIS observations until 2004

auto0601/ : code for THEMIS observations since 2005

dalsa/ : code for VTT docdpsm/ dpsm_form/

Results will be in auto0503/ or auto0601/

+ 0503

VTT processing

Similar algorithm but manual (no msdpauto, no sequence.par)

editing of dalsa.pro, ms.par (file time …)

Start ms1 directly

In autoXXXX/

The msdpauto (IDL) will : create directories for the results

dir_date_seqnum_camnum/ copy the code in it go to this directory build the ms.par start the fortran code put the results thereCopy the resulting g* file (geometry) in the

upper directory in case useful for another sequence

Goto /data4/groupX/ or /baie1/groupX/ Goto the autoxxxx/ corresponding to the

sequence Make the sequence.par if not provided

manually (see example in /data3/DPSMWS/sequence_model.par)

automatic with the BASS2000 interface To get bs + camera number : edit fits file

x=mrdfits(/data…./*.fts’,ran=[0,0]) (WARNING 2004 : 2 cameras / line)

Depending on the case : start msdpauto, « /dir_data/ » start msdpauto, « /dir_data/ », /no_fort, modify

ms.par then start ms1 other procedure if indicated

File vizualization

How to process a sequence

File vizualisation

IDL : x=readmsdp(‘file’) (q* or p* file) check the image order in the file tvscl,x(*,*,i) ; plot… (cut) etc.

When using the BASS2000 interface : choose q or p, and « Voir DEP » : equivalent to tvscl

Postscript files : kghostview file.ps ; or « Voir PS » in the BASS2000 interface

Rep auto

Makesequence.par

msdpauto

Edit sequence.par if modificationIf modif ms.par

Categories of sequences

Good sequences (THEMIS, VTT) Sequences with interruption during

the processing Sequences providing maps or

profiles showing problems (visual verification)

For each sequence

Date, sequence number Line(s) Volume Time to process (for 100% CPU time on

our machine) Problem if any, illustration before / after

To reduce processing time after transformation b3 to individual files

: tob1, tob2 in ms.par ; or nob1, nob2 (msdpauto should have been started with /no_fort)

stop at « q » step (qp = 0 in sequence.par) for quicklook only (no profile)

keep only one stokes if several are observed (qv, nq in sequence.par) ; see SEQ_STOKES in fits header

Beam exchange : nqseul = 7, 8, 9 in ms.par (otherwise 0)

Sequence : #1, 9 aug 2004Line: 5896Volume : 98 MoSeq_stokes : +VProcessing time : <16 minNo problem

Sequence : #1, 20 oct 2003Line: 5896 & 8542Volume : 2.2 GoSeq_stokes : +Q+U+VProcessing time : <28 min / lineNo problemUse nq=3, qv=3

Sequence : #3, 20 oct 2003Line: 8542 (cam b01)Volume : 5.9 Go (2 lines)Seq_stokes : +Q+U+VProcessing time : <1h30 No problemUse nq=3, qv=3

Sequence : #7, 31 oct 2003Line: 5896 & 8542Volume : 1.7 GoSeq_stokes : +V+Q+UProcessing time : <35 min / lineNo problem Use qp=3, qv=1

Sequence : #6, 25 oct 2002Line: 5896 (cam b02)Volume : 3.9 GoSeq_stokes : +VProcessing time : <1h20No problem

Sequence : #3, 25 oct 2002Line: 5896Volume : 1.6 GoSeq_stokes : +VProcessing time : <1h30No problem ; illustration dt=0,60Use tob1, tob2

dt=60

dt=0

Sequence : #7, 6 oct 2005Line : 5896Volume : 660 MoSeq_stokes : +VProcessing time : 20 mindt/ntmax problem + intert

dt=60 in sequence.par

dt=0 in sequence.par ; or ntmax=6, intert=0 in ms.par

Sequence : 17 sept 2005, VTTLine: Halpha, no polarVolume : 2.7 GoNo problem

Sequence : #2, 31 jul 2004Line : 5896Volume : 300 MoSeq_stokes : +VProcessing time : 16 minmilgeo problem, must use seq #5 g* fileIt is also possible to impose #seq calib in sequence.par

Seq 5, 9:26 Seq 2, 7:57

Sequence : #12, 2 aug 2004Processing time : 20 minmilgeo problem, illustrate very large value of milgeo

Sequence : #14, 31 jul 2004Processing time : 15 minillustration bad data

Sequence : #13, 13 oct 2003Processing time : 18 minillustration bad data

Sequence : #4, 14 oct 2003Processing time : 1h05illustration bad data

Sequence : #3, 15 oct 2003Processing time : 48 minillustration bad data

Sequence : #11, 10 aug 2004Processing time : 23 minKeyword problem : polarization or not ?

Sequence : #3, 15 oct 2004Line : 5896Volume : 287 MoSeq_stokes : +V-VProcessing time : 1h03 min for whole seqinveri problem +xfirst + illustration tob1/tob2Use ilisqp=3, jlisqp=3 or more to smooth

inveri=0

inveri=1

inveri=1, tob1=15123467, tob2=15123608

Sequence : #3, 16 oct 2002Line : 5896Volume : 1.3 GoSeq_stokes : +VProcessing time : 2h10 min whole scanxfirst problem + norma + illustration tob1/tob2

xfirst=0

xfirst=1

xfirst=1,norma=0199

Sequence : #11, 28 oct 2002Line : 5896Volume : 1.1 GoSeq_stokes : +VProcessing time : ?Small maps with scan problems

dt=60, xfirst=0

dt=0, xfirst=0

dt=0, xfirst=1