[Frost] gsim problem

Sat Sep 17 20:23:11 EDT 2011

Thanks, Franz!  ---BGR

Professor Barry G. Ritchie
Department of Physics
Arizona State University
Tempe, AZ  85287-1504

Telephone: (480) 965-4707

________________________________

From: frost-bounces at jlab.org on behalf of Franz Klein
Sent: Sat 9/17/2011 8:51 AM
To: Michael Dugger
Cc: frost at jlab.org
Subject: Re: [Frost] gsim problem

Mike et al,

after dumping calibration constants from caldb and Map and printing lots
of time-based tracking variables ..., I found that at the end it were
simply my latest changes to 'sc_init.c' where I added ranges for the
calibration constants because gamecock (sc calibration program) gave lots
of bad fit results, which caused user_ana to crash on real data.
After changing the ranges to include zeros where gsim requires it, it's
working:
------- track statistics at the end of recsis/user_ana:
TRK_STATI  L: Cut#  Description              Value    Passed     %
...
TRK_STATI  L: 8  ana_fit:   Nevt FF1 chi2 <   50.0     57    57.0   70.4
TRK_STATI  L: 9  ana_fit:   Nevt in TBT >      0.0     57    57.0   70.4

Have fun ... (I had enough for the weekend)
Franz

On Fri, 16 Sep 2011, Franz Klein wrote:

>
> Folks,
> at CLAS the code for reconstruction and simulation (gsim-gpp-recsis chain) is
> not really the same: the simulation code generates some design response
> (ADC,TDC) for the detectors (which are smeared a little in gpp), thus MC
> reconstruction requires different calibration constants (and DC
> drift-distance function).
> Please follow (at least in part) the parameter settings in
> packages/tests/frost (gsim_frost.ffread and recinit_gsim.tcl),
> e.g. setting beam_type to 4 results in using the c++ code for reconstruction
> ... but it seems that the old fortran code works better;
> e.g. only my old fortran version of the left-right ambiguity fit seems to be
> working properly ('set trk_lramb -1; in recinit_gsim.tcl .... I have to check
> that some time), but I have not figured out why running at jlab machines
> requires to set the chi^2 cut for the time-based fit to 'set trk_tbtfit_chi2
> 1000' (it works with a cut at 50.0 nicely at cua and at home);
> e.g. the nasty warning in recsis/user_ana:
> RECKOUT    W: -----No DATA OUTPUT STREAM found----
> is an artifact when compiling with the bosio (c-code) package instead of the
> old BOS/FPACK (fortran). I corrected the routine such that it does not
> complain if the BOS output file was opened by a tcl script procedure.
>
> Mike's problem was rather simple to fix after finding it: since Qweak needs
> high rates of helicity flipping 'class 15' in the HEAD bank is reserved for
> events with dumps of helicity scalers. I changed the code in recsis to
> reconstruct events of 'class 15' if they are Monte Carlo generated events
> (the CLAS libraries in /home/clasg9 are recompiled).
> After this change and adding/changing in recinit_gsim.tcl the flags:
> set trk_lramb -1;
> set trk_tbtfit_chi2 1000;
> I got the 100 events through user_ana:
> TRK_STATI  L: Cut#  Description               Value    Passed     %
> TRK_STATI  L: 1  trk_evnt:  Ntrig (Nevt) >     0.0      100   100.0
> TRK_STATI  L: 2  ana_segm:  Nevt w. segm =     5.0       84    84.0
> TRK_STATI  L: 3  trk_link:  Nevt w. link =     5.0       83    83.0  100.0
> TRK_STATI  L: 4  ana_prfit: Nevt PR chi2 <    20.0       83    83.0  100.0
> TRK_STATI  L: 5  SEB_eid:   Nevt SEB     =     0.0       59    59.0   71.1
> TRK_STATI  L: 6  trk_dcam:  Nevt w. LRseg=     3.0       59    59.0   71.1
> TRK_STATI  L: 7  trk_dcam:  Nevt w. LRseg=     4.0       59    59.0   71.1
> TRK_STATI  L: 8  ana_fit:   Nevt FF1 chi2 < 1000.0       50    50.0   60.2
> TRK_STATI  L: 9  ana_fit:   Nevt in TBT >      0.0       50    50.0   60.2
>
> I updated recinit_gsim.tcl and mytest_frost.sh (there were some flags in
> mytest_frost.sh which were not correctly treated in the script).
> Please check out or update: cvs co tests/frost
>
> Good luck
> Franz
>
> ps. still have to figure out which DC Sigma's are far too small (=blowing up
> the chi^2)... maybe tomorrow.
>
> On Fri, 16 Sep 2011, Eugene Pasyuk wrote:
>
>> Mike,
>>
>> I see in the log beam type 1. I think it should be 4 to get reconstruction
>> with new start counter. I also see that RF reconstruction is initialized.
>> Tagger and RF reconstruction should be turned off in cooking of gsim data.
>>
>> -Eugene
>>
>> On 9/15/11 14:58 , Michael Dugger wrote:
>>>
>>> Hi,
>>>
>>> After making the changes Franz suggested, I'm still having problems
>>> getting the gsim bos file to cook.
>>>
>>> I have attached my output of the cooking from a gsim file (out.txt)
>>>
>>> Patrick found something in the out.txt file that might be of interest:
>>>
>>> RECKOUT    W: -----No DATA OUTPUT STREAM found----
>>>
>>> Any comments or suggestions are appreciated.
>>>
>>> Thanks,
>>> Michael
>>>
>>>
>>> _______________________________________________
>>> Frost mailing list
>>> Frost at jlab.org
>>> https://mailman.jlab.org/mailman/listinfo/frost
>>
>
> ===============================================================
>                  Franz J. Klein, Associate Professor
>                  CUA, Department of Physics
>                  Washington, DC 20064
>  office: Hannan Hall 206          phone: 202-319-6190
>  or: Jefferson Lab,CC F-243       phone: 757-269-6672
> ---------------------------------------------------------------
>

===============================================================
                   Franz J. Klein, Associate Professor
                   CUA, Department of Physics
                   Washington, DC 20064
   office: Hannan Hall 206          phone: 202-319-6190
   or: Jefferson Lab,CC F-243       phone: 757-269-6672
---------------------------------------------------------------
_______________________________________________
Frost mailing list
Frost at jlab.org
https://mailman.jlab.org/mailman/listinfo/frost