[Frost] gsim and rootbeer issues (fwd)

[Michael Dugger]

Hi,

Below is a response from Ken regarding access to single sector banks 
(e.g. MCTK and MCVX) within the rootbeer framework:

---------- Forwarded message ----------
Date: Wed, 21 Sep 2011 09:23:32 +0100
From: Ken Livingston <Kenneth.Livingston at glasgow.ac.uk>
To: Michael Dugger <dugger at jlab.org>
Subject: Re: [Frost] gsim and rootbeer issues

Hi Micheal,
The BOS banks can be either single sector or multi sector.
The problem is that there's no way of telling from the data, or from the BOS 
ddl files.
You just have to know.

So, for safety, unless specified in the list of banks in 
$ROOTBEER/dat/singles.dat, rootbeer treats all banks as multi sector.

There are 2 ways round this for single sector banks. The quickest and easiest 
is option 2.

1. You can specify the bank in  $ROOTBEER/dat/singles.dat and rebuild by
doing a make in the $ROOTBEER directory. I believe the Makefile is good enough 
to cope with this.
I've attached singles.dat which Craig Paterson used (including simulation 
related banks).

2. You can use a BANK as a multi sector BANK, but only look at the first 
sector.
eg. instead of
this:  hVertMCVX->Fill(MCVX[0].z)

try this:    hVertMCVX->Fill(MCVX[0][0].z)

ie. if singles.dat is empty all banks can be addressed as 
BANK[sectorNo][hitNo], but for the singles, sectorNo is 0.
I wanted to make things conceptually easier for analysis, and get away from the 
clumsy notation of other bank handling software, that's why I also added the 
BANK[hitNo] notation. Not sure, in retrospect, it was such a such a wise move.

Let me know if this works.

Cheers,
Ken

ps. tried to send this to FROST list, but its bouncing my attempts for some 
reason. Maybe you could forward it.



On 09/20/2011 05:08 PM, Michael Dugger wrote:
> Hi,
> 
> GSIM issue:
> GSIM is smearing the target vertex wrong. The event generator I use sets
> all of my reaction vertexes at 0,0,0 but after gsim the target is smeared
> and the center ends up being at approximately + 4cm.
> 
> In the gsim_frost.ffread file, the target center is set to zero:
> TGPOS 0.0 0.0 0.0
> 
> The start counter offset:
> STZOFF 4.0
> is the same as the offset I am seeing for the vertexes. Is this just a
> coincidence?
> 
> The way I determine that the target is offset by 4 cm is by looking at
> vertex from MVRT and TBTR (passed through GPID). I am having difficulties
> plotting directly from MCVX (more on this later). I spot checked the MCVX
> bank from the event generator to make sure the vertex was always at zero.
> When I bosdump the output from gsim_bat (on the Jlab CUE) the
> smeared vertex from MCVX appears to be consistently shifted downstream
> consistent with the shift being about 4 cm.
> 
> ROOTBEER issue:
> When I print out the MCTK_NH or MCVX_NH variables I always get 0. I have
> checked that the root file I process has the MCTK and MCVX banks by
> looking at the output of the bankdump command. When I try to compile the
> code accessing individual members of the banks (MCTK and MCVX), I get
> errors like:
> /home/dugger/g9Ana/test/testDST.C:348: error: request for member z in *
> MCVX, which is of non-class type MCVX_t*
> 
> The code I am using has for the bank definitions:
> char 
> *mybanks[]={"RUNC","GPID","MVRT","BEAM","TAGR","TGBI","null","MCVX","MCTK"};
> 
> And I try to access the z-vertex of the MCVX bank like this:
> hVertMCVX->Fill(MCVX[0].z);
> 
> Should I be accessing the z-vertex from MCVX in a different way?
> 
> Thanks for your time.
> 
> Sincerely,
> Michael
> 
> _______________________________________________
> Frost mailing list
> Frost at jlab.org
> https://mailman.jlab.org/mailman/listinfo/frost


-------------- next part --------------
//list of known single sector banks single
HEAD
TESC
GP_X
GP_Y
CALL
PRIM
STN0
STN1
ST
TGBI
CL01
DSTC
CCRC
HBTR
HBER
HBID
ECPI
ECHB
TBTR
TBID
TBER
EVNT
DCPB
ECPB
SCPB
STPB
TGPB
HEVT
LCPB
CCPB
TAGE
TAGT
TAGR
TAGM
HISS
PART
STR
MVRT
MCVX
GPID
GPIE
TGTR
TGTL
REF
RC26
RF