Re: [Brahms-dev-l] global tracking

From: Ian Bearden <bearden@nbi.dk>
Date: Mon Jul 04 2005 - 04:53:05 EDT
I would suggest that we start by redoing a couple of runs where the  
"efficiency" is low, and a couple where it is high.
Let us redo ltr, gtr, dst (at the same time, and thus, presumably,  
with exactly the same software) and see how it looks.
The only guess I have as to why the number of tracks falls with  
version number is that (maybe?) someone has
"improved" some cuts.
Cheers,
Ian
On 01/07/2005, at 18:54, flemming videbaek wrote:

> During the ewednesday meeting there was a discussion on the a  
> behaviour observed by Catalin
> see
> http://www.nbi.dk/~ristea/QM05/
>
> See the t4t5/ffs etc..
>
> Pawel pointed out that (a little while back too) that the DC  
> enhancement was not always on.
> the two jumps at 10300 and 10484 does not seem to be correlated with
> - anything noted in elog (basically change of down scales during a  
> store.
> - calibrations for any sub system.
> - looking at overall rates of tracks i.e. global tracks/trigger 2  
> there is a slight change with runs
> 299,300,301 -> 0.293,0.317,0.33, - this could be related reducing  
> backgrounds. The H2 rates drops quickly throughout these runs.
> Thus no solutions
>
>
> By coincident I looked at run 10384/10385. This was interesting  
> since it has 3 sets of gtr's
> stting around. Apart from this it was also rerun recently on June  
> 22 (coincidently not using the submitGtr.perl script ) with DC  
> enhancement on.
>
> the truely remarkel fact is that the number of good tracks getting  
> out of the globals tracking
> drops as more reconstruction is doen
>
> version     T4T5    T3T4   FFS  T24  GlobalTracks
>      6            320      340    2354   528     132
>      5            328      319    2359   527     167
>      4            328      319    2359   527     188
>
> How can this be the case !!
> I find this exceedingly worrisome. Since the numbers of overall  
> track in individual combinations
> seems ok, how come that tracks are lost with time (of course they  
> may not have been good tracks, but it should be looked at.
>
>
> flemming
>
>
>
>
>
> ----------------------------------------------------------------
> Flemming Videbaek
> Physics Department
> Brookhaven National Laboratory
>
> e-mail: videbaek@bnl.gov
> phone: 631-344-4106
> _______________________________________________
> Brahms-dev-l mailing list
> Brahms-dev-l@lists.bnl.gov
> http://lists.bnl.gov/mailman/listinfo/brahms-dev-l
>




_______________________________________________
Brahms-dev-l mailing list
Brahms-dev-l@lists.bnl.gov
http://lists.bnl.gov/mailman/listinfo/brahms-dev-l
Received on Mon Jul 4 04:53:22 2005

This archive was generated by hypermail 2.1.8 : Mon Jul 04 2005 - 04:53:32 EDT