how to run a standalone event production including showering, and hadronisation

How is installed pythia8 in your setup?
I guess that this is part of the mg5_configuration_LCG102_x86_64-centos9-gcc11-opt.txt
file that you copy. Is that correct?
But I see that you have to install mg5amc_py8_interface, that sounds weird (but possible if they install pythia8 externally and do not care about running at LO).

my suggestion would be:
echo "generate p p > z j j QCD=0" > cmd
echo "output zjj_ewk" >> cmd
echo "launch zjj_ewk" >> cmd
echo "shower=PY8" >> cmd
./bin/mg5_aMC cmd

Obviously it will only work if the pythia8 is corectly setup. (which is needed anyway to have the install of the mg5amc_py8_interface working).

Now note that we had a recent bug for pythia when running it in single core.
So if you see an error with "unset" this means that such patch is not life yet.

> Incidentally, I have a generic question about config files.

The config file are read in the following order:
when running from ./bin/mg5_aMC executable:
0) value default (hardcoded in the code)
1) system-wide config file (need environment variable MADGRAPH_BASE)
2) home config file (need environment variable HOME)
3) input/mg5_configuration.txt

When running from "process" directory:
0) value default (hardcoded in the code)
1) system-wide config file (need environment variable MADGRAPH_BASE)
2) home config file (need environment variable HOME)
3) Cards/me5_configuration.txt
4) input/mg5_configuration.txt (need that of the previous step defines mg5_dir=)
5) Cards/me5_configuration.txt (rerun only if "mg5_dir" was set)

So yes, if you specify something in Cards/me5_configuration.txt this has always priority.
If you do not, then we try to use MG5_aMC_v3_5_4/input/mg5_configuration.txt
but this is possible only if a mg5_dir is specified in one configuration file (and point to a valid madgraph directory)

Cheers,

Olivier

Dear Oliver,

thanks a lot for the quick reaction.
Concerning your question, the pythia8, fastjet, HepMC, Delphes programs come from an LHC environment, which is set up with the command:

source /cvmfs/sft.cern.ch/lcg/views/setupViews.sh LCG_102 x86_64-centos9-gcc11-opt

so I expect the installation to be sound.
The file mg5_configuration_LCG102_x86_64-centos9-gcc11-opt.txt was created by me to tell Madgraph where all packages are.

I will try following your suggestions and check for all the environment variables and come back to you.

Thanks!

pietro

Dear Olivier,

I gave it a try and I still have a couple of questions I'd like to ask:

- when I use the instruction: shower=PY8 the shower gets deactivated, while if I do nothing it's there by default, which I find strange, how do I make sense of it?

it seems that the information present in mg5_configuration.txt does not get properly parsed, evidence 1, since I set it up NOT to run on condor, while pythia jobs get submitted to condor, how can I prevent it?

- it seems that the information present in mg5_configuration.txt does not get properly parsed, evidence 2, since the hepMC files are not saved because the MG5aMC_PY8_interface cannot be found, even if it's set. This is the error I get:

/pool/condor/dir_2863694/run_PY8.sh: line 2: ./MG5aMC_PY8_interface: No such file or directory

while this is the setting in the file:

mg5_configuration.txt:mg5amc_py8_interface_path = /afs/cern.ch/work/g/govoni/2024-VBS-tagger/LCG/MG5_aMC_v3_5_4/HEPTools/MG5aMC_PY8_interface #

Do I need to manually copy the file in the process folder?

Best regards,

pietro

Dear Olivier,

I did manage to solve the condor problem, I found a bug in my code.

pietro

So everything is fixed now, or you still need help on something?

Cheers,

Olivier

Hello Oliver,

yes, thanks, this thread has been very useful.
I am still left with some open points in my flow but as of now I can produce events, which is what matters :)
In case of more doubts I will open another ticket.

Have a nice day,

pietro