Date: Fri, 20 Sep 2002 22:34:12 +0200 (MET DST)
From: Lukas Phaf <lphaf@nikhef.nl>
To: Volker Buescher <buescher@fnal.gov>
Cc: Terry Wyatt <wyatt@hep.man.ac.uk>, Jonathan Hays <haysjm@fnal.gov>,
     lphaf@fnal.gov
Subject: Re: v9.0 filter certification

Volker asked my to run trigsim p12 with triggerlist v8.2 over
a data run taken with the same triggerlist. This to make it possible
to compare p12(offline) with p11(online).
I ran over 3 raw data files of run 163172. The result is a root
tuple with both online and offline branches.
I see slight differences in all physics tools, slightly different
numbers of tracks/electrons/muon/jets/taus found, and differences
in the distributions.

The root tuples is:
/work/diemen-clued0/lphaf/p12.02/trigsim_data/3FILE/L1L2L3global.root

The level3.sim files used is:
/work/diemen-clued0/lphaf/p12.02/trigsim_data/3FILE/level3.sim

Steps needed to set up trigsim to run like this.

I use the test version of d0tools (setup d0tools -t),
and p12.02 version of d0trigsim.

setup D0RunII p12.02.00
setup d0cvs
newrel p12.02.00 trigsim
cd trigsim
addpkg d0trigsim
addpkg l3fanalyze
setup d0tools -t

I use the trigger list found on
http://www-d0.fnal.gov/computing/trigsim/general/triglists.html
(v 8.20), I used the "coor" version, and modified according too:
http://www-d0.fnal.gov/computing/trigsim/release/p12.html

So:
(In the trigsim directory)
mkdir trigsim
cd trigsim
get the tar file, and unpack it. Edit (level2.sim and) level3.sim as
shown on the webpage.

cd ..

To set d0trigsim up so that is produces both online and offline
blocks, make the following modifications:

edit d0trigsim/rcp/runD0TrigSim_l3only_data.rcp
The Packages string should read:
string Packages = "init coor l3_an_on tsiml3 l3_an_off event_info"

l3_an RCP lines should read:
RCP l3_an_on =<l3fanalyze L3Analyze_auto_online>
RCP l3_an_off =<l3fanalyze L3Analyze_auto_offline>


To make d0trigsim use the L2 online decisions to decide which filters
to run (default is to run all filters):
In the trigsim/level3.sim files, the SRDirective line should read:

trigger_list 0 SRDirective (useL2=yes)|

(where originally it has useL1=no.
WARNING: This only works if the trigger bit numbers are the same
in the level3.sim and the actual run. Look at the end of the level3.sim
file, and the runs database: http://d0db.fnal.gov/run/runQuery.html


In d0trigsim/rcp/D0TrigSimInit_data.rcp, remove run_config from
the Packages string.

That should do it, run I run with:
runD0TrigSim -filelist=run163172.dat -format=data -localrcp -l3only 
-localfwkrcp -name=3FILE -batch -cput=72:00:00 -mem=400mb 
-jobname=run163172_3

Lukas Phaf

P.S. There is an issue with a HepTuple bug. There is a fix, but
a quick fix that worked for me is turning off the unpacker examiners.
There examiner list is generated automatically by running d0tools,
so a little dirty trick is needed:

Create the complete examiner list:
cd trigsim
ParseToolList level3.sim toollist.txt
cp toollist.txt toollist.stripped.txt
edit toollist.stripped.txt, remove the lines:
L3TCalUnp CAL_UNP
L3TCFTUnpack CFTUnp
L3TSmtUnpack SmtUnp

To use this stripped version in your run:
cd ..
edit l3fanalyze/rcp/L3Analyze_auto_offline.rcp
and l3fanalyze/rcp/L3Analyze_auto_online.rcp,
at the bottom, replace toollist.txt by toollist.stripped.txt