"Productivity" Tutorial

Dear Bjoern,

The use of the “Template” directory is the method of MG4 which was indeed still supported for retro-compatibility up to MG5.1.5.11.
Since MG5 1.5.5 (not sure of the exact version number) it was clearly indicate that this retro-compatible mode would be dropped
and indication on how to move to the more modern way was indicated on the screen.

> Looking up tutorials one find only trivial examples using the MG5 shell, unfortunately not that useful if one want to produce large datasets and/or a range of parameters.

That is actually not True most of my tutorial (if not all) have one exercise showing how to
1) bypass the shell interface
2) perform a scan on one parameter
see for example:
https://cp3.irmp.ucl.ac.be/projects/madgraph/attachment/wiki/MGTalks/13_06_10_tutomg_tasi.pdf
(exercise 5)

> * How can I run MG5_aMC in an automated fashion? What and where are the equivalents of process and params files, are the Card files in 'Template/Common/Cards/' still the relevant ones. Can I still use 'Template' as base of all processes?

In MG4, the correct way to generate a process was:

cp -r Template PATH
cd PATH
EDIT ./Cards/proc_card.dat
./bin/newprocess
EDIT ./Cards/run_card.dat
EDIT ./Cards/param_card.dat
./bin/generate_events OPTIONS

Some comment
# with MG5.1.X, you can do the same with either proc_card.dat or proc_card_mg5.dat (the only one present in the Template of that version)

Since MG5.1.X the official way is (and only way since MG5.2.X)
EDIT proc_card.dat
./bin/mg5_aMC proc_card.dat

cd PATH # PATH output is define in proc_card.dat
EDIT ./Cards/run_card.dat
EDIT ./Cards/param_card.dat
./bin/generate_events OPTIONS

Note that
- the file proc_card.dat has the same syntax as of the proc_card_mg5.dat (it can also be a v4 format for the proc_card actually).
- you can rename that file, or start from a fresh file if you want to, the name of the file is not important.
- PATH is define in proc_card.dat with the command “output PATH”
- a lot of script of MG4 present in the bin directory (survey/refine/pythia/combine/multi_run) have been moved to be part of the ./bin/madevent shell.
  (since MG5.1.4.X) Those can be scriptable as well (one exa.
- The edition of the card is scriptable inside MG (since MG5.1.5.X) and you can therefore make scan easily.
one example of scan (+use of one of those removed script) is
1) edit a file lest’s call it “cmd.txt"
generate_events
 set mt 175
        set decay 6 Auto
multi_run 3
       set nevents 50000
       set mt 180
       set decay 6 Auto
2) run the code ./bin/madevent cmd.txt

> it seems to be intellectually lazy to mourn the passing of the old way to run it. But the absence of a consistent and prominent manual make it appear that there is a steep learning curve consisting mainly out of trial & error. Perhaps you can help?

All of this is actually explained in our two last paper.
MG5 paper: arXiv:1106.0522
MG5_aMC paper: arXiv:1405.0301
we also have most of the above information include in one of our FAQ:
https://answers.launchpad.net/mg5amcnlo/+faq/2186
The manual is actually the code itself (with the help command) and this launchpad website.

Cheers,

Olivier

On 11 Feb 2015, at 12:11, Bjoern Penning <email address hidden> wrote:

> New question #261959 on MadGraph5_aMC@NLO:
> https://answers.launchpad.net/mg5amcnlo/+question/261959
>
> Dear Madgraph Authors
>
> I have made in the past feeble attempts to move on from the 'old' way to run MG to the new MG5_aMC way and every time I bounce right back because the changes introduced to the system are different enough to break productivity and yet there are enough similarities that the entire ordeal becomes frustrating.
> Also there is a need for more consistent documentation aiming for 'productivity' use. Many of my colleagues (exp. & theory) have had the similar experiences.
>
> Typically we experimentalists receive some UFO files from our theory colleagues, plug it into an copy of the old 'Template' directory and run generation with 'bin/newprocess' and 'bin/generate_events' with some changes to params, run files etc.
>
> Trying to move to MG5_aMC there we see the changed structure in Template that now contains a Common directory with some 'bin' and some card files, but the crucial commands (newprocess, generate_events) and also the crucial process and params files are not there(?!).
>
> Looking up tutorials one find only trivial examples using the MG5 shell, unfortunately not that useful if one want to produce large datasets and/or a range of parameters.
>
> So, a bunch of questions and one suggestion:
>
> * How can I run MG5_aMC in an automated fashion? What and where are the equivalents of process and params files, are the Card files in 'Template/Common/Cards/' still the relevant ones. Can I still use 'Template' as base of all processes?
>
> it seems to be intellectually lazy to mourn the passing of the old way to run it. But the absence of a consistent and prominent manual make it appear that there is a steep learning curve consisting mainly out of trial & error. Perhaps you can help?
>
> Thank you, Best
> Bjoern
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Dear Olivier

Thank you very much for the answer, I think I was using a way too outdated workflow and then got completely thrown of. With the information you provided I was able to get it running quickly and agree. It's a way better set up. I could have saved a lot of time changing earlier.

Thanks
 Bjoern