[Clas_offline] Fwd: Generated events in the PART bank (PART bank sector 0) after user_ana.

Tue May 20 14:25:53 EDT 2014

Has this always been the case?  If so, is it generally known?  If not, this is a serious issue which should be addressed, since we need to know how many events we're simulating if we want to calculate our acceptances correctly.

John

--
John Price
Professor and Chair, CSUDH Department of Physics
Co-Director, Office of Undergraduate Research, Scholarship, and Creative Activity
(310) 243-3403

On May 20, 2014, at 11:07, "Haiyun Lu" <hlu at jlab.org<mailto:hlu at jlab.org>> wrote:

Hi Nick,

I believe it has nothing to do with gsim and fsgen is the cause. Did you check how many events you got out of fsgen? I bet it's 71 events. You better check the evt file directly instead of the log file. To my knowledge, Fsgen makes 100 loops to generate events when you ask for 100 events. However, some events are discarded due to some reason, like it's not physical. Fsgen will NOT make up for these events. Therefore you only get a portion of what you ask for.

Haiyun

On 05/20/2014 01:41 PM, Johann Goetz wrote:
Hi All,
Nick Compton is having what looks like a very similar issue to Markov's (see below). Was this ever resolved?
-- Johann<https://sites.google.com/site/theodoregoetz/>

########## begin message from Nick Compton #########

when I run a Monte Carlo simulation say of 100 events, I only get 71 back in the root tree. This may be what I expect from a detector simulation depending on the reaction, except for the fact that in the mc bank I am also only getting 71 events. This seems odd as I would not expect the number in the mc bank to be the same as the events bank. Do you know if there is an option I have to use with GSIM or in another part of the process to have the mc bank work properly? Please let me know of any insight you might have.

I am using fsgen to generate .evt files which according to the .log file produced 100 events. The information is stored in the directory '/volatile/clas/clasg10/ncompton'.

Thanks,

Nick

########### begin thread of emails associated with Markov's problems ########

---------- Forwarded message ----------
From: Michael C. Kunkel <mkunkel at jlab.org<mailto:mkunkel at jlab.org>>
Date: Fri, May 16, 2014 at 9:20 AM
Subject: Re: [Clas_offline] Generated events in the PART bank (PART bank sector 0) after user_ana.
To: nick markov <markov at jlab.org<mailto:markov at jlab.org>>, KiJun Park <parkkj at jlab.org<mailto:parkkj at jlab.org>>
Cc: clas offline <clas_offline at jlab.org<mailto:clas_offline at jlab.org>>

Greetings,

From what I recall, the STABLE build of RECSIS/user_ana will not output PART 0. However, I remember Stepan having a build that does output PART 0. Maybe you can compare the source codes and figure out why.

BR
MK

On 5/16/14 9:13 AM, nick markov wrote:
Hi,
thank you very much for your respond. The tcl file which I use is :
    source /group/clas/builds/test3/src/clas6-trunk/reconstruction/recsis/recsis_proc.tcl;
    turnoff ALL;
    global_section off;
    turnon seb trk cc tof egn lac user pid;
    inputfile gpp_out.bos;
    setc chist_filename   recsis.hbook;
    setc log_file_name   out.log;
    outputfile test.recsis PROC1 2047;
    set dc_xvst_choice     0;
    set torus_current      2250;
    set mini_torus_current 6000;
    set poltarget_current  0;
    setc outbanknames(1) "PARTEVNTECPBHEADDCPBTBIDTBTRTDPLTBLACCPBTRGSSCPBTGBI";

    setc prlink_file_name "prlink_NEW_60_75.bos";
    setc bfield_file_name "bgrid_T67to33.fpk";

    set lseb_nt_do  -1;
    set lall_nt_do  -1;
    set lscr_nt_do  -1;
    set lseb_hist   -1;
    set lseb_h_do   -1;
    set lmon_hist   -1;
    set ltrk_h_do   -1;
    set legn_h_do   -1;
    set ltof_h_do   -1;
    set lec1_h_do   -1;
    set lfec_hist   -1;
    set lfec_h_do   -1;
    set lpart_nt_do -1;
    fpack "timestop -9999999999";
    set lscat $false;
    set ldisplay_all $false;
    go
    exit_pend;
Best  regards, Nick.

On May 16, 2014, at 9:08 AM, KiJun Park <parkkj at jlab.org<mailto:parkkj at jlab.org>> wrote:

Hello,

Interesting. Is this matter of creating nPRT in Ntuple after RECSIS ?
For example,
*      1   * I*4  *         * [0,20]       * PART     * nPRT
*      2   * I*4  *         *              * PART     * pidPART(nPRT)
These are empty ?
If so, I believed that it is related with set variable.
I wonder whether Nick can show his tcl file.

Kijun

----- Original Message -----
From: "Stepan Stepanyan" <stepanya at jlab.org<mailto:stepanya at jlab.org>>
To: "clas offline" <clas_offline at jlab.org<mailto:clas_offline at jlab.org>>
Sent: Friday, May 16, 2014 8:30:10 AM
Subject: Re: [Clas_offline] Generated events in the PART bank (PART bank sector 0) after user_ana.

Johann,

I am not familiar with how PART is handled in user_ana, it is in pid package I am sure.
But it is more likely that PART 0 is not deleted but it is just is not written out.

Stepan

On 5/16/14 7:47 AM, Johann Goetz wrote:

Does anyone know why user_ana should actively delete the PART bank sector 0 from the data? This sounds like a bug. In principle, gsim will digitize PART just the same it does for MCTK/VX. The low-level banks produced, and used by the reconstruction packages, will be the same.

In short, I believe the solution to Nick's problem is to NOT DROP the PART bank sector 0 during reconstruction (user_ana). Can someone tell him how to do this using user_ana?

-- Johann

On Tue, May 13, 2014 at 3:47 PM, Gagik Gavalian < gavalian at jlab.org<mailto:gavalian at jlab.org> > wrote:

Hi Nick,

The PART 0, is used when you're running simulation and then reconstructing with a1c, which keeps both PART 0 (simulated bank) and PART 1 reconstructed bank.
If you're using user_ana to for reconstruction, you should write your simulation into MCTK and MCVX banks, then run GSIM/RECSIS chain on it.

Gagik

On May 13, 2014, at 3:38 PM, nick markov < markov at jlab.org<mailto:markov at jlab.org> > wrote:

user_ana

_______________________________________________
Clas_offline mailing list
Clas_offline at jlab.org<mailto:Clas_offline at jlab.org>
https://mailman.jlab.org/mailman/listinfo/clas_offline

_______________________________________________
Clas_offline mailing list Clas_offline at jlab.org<mailto:Clas_offline at jlab.org> https://mailman.jlab.org/mailman/listinfo/clas_offline

_______________________________________________
Clas_offline mailing list
Clas_offline at jlab.org<mailto:Clas_offline at jlab.org>
https://mailman.jlab.org/mailman/listinfo/clas_offline

--
---------------------------------------******
Dr. Kijun Park
Physics Dept. Old Dominion University
Thomas Jefferson National Accelerator Facility
Suit 5 Room 12_1/A102
12000 Jefferson Ave.
Newport News VA 23606
Phone 757-269-6989<tel:757-269-6989>

_______________________________________________
Clas_offline mailing list
Clas_offline at jlab.org<mailto:Clas_offline at jlab.org>
https://mailman.jlab.org/mailman/listinfo/clas_offline

_______________________________________________
Clas_offline mailing list
Clas_offline at jlab.org<mailto:Clas_offline at jlab.org>
https://mailman.jlab.org/mailman/listinfo/clas_offline

_______________________________________________
Clas_offline mailing list
Clas_offline at jlab.org<mailto:Clas_offline at jlab.org>
https://mailman.jlab.org/mailman/listinfo/clas_offline

_______________________________________________
Clas_offline mailing list
Clas_offline at jlab.org<mailto:Clas_offline at jlab.org>
https://mailman.jlab.org/mailman/listinfo/clas_offline

_______________________________________________
Clas_offline mailing list
Clas_offline at jlab.org<mailto:Clas_offline at jlab.org>
https://mailman.jlab.org/mailman/listinfo/clas_offline
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://mailman.jlab.org/pipermail/clas_offline/attachments/20140520/1d617408/attachment-0001.html 