MadSpin stuck during "decay events", many "check" processes running

Hi,

This is part of the limitation of the code :
see here:
https://amcatnlo.web.cern.ch/amcatnlo/

To handle such process in five flavor you need to remove the third top resonance, and you need to look at MADSTR plugin for that.

Cheers,

Olivier

Hi Olivier,

thanks, I see... this is one of our validation jobs, strange that it worked in the past.
Running the process in 4fs would work then?

Cheers,
Hannes

Yes four flavor should not have issue.

This bieing said, the limitation is not one of madspin but one of the NLO computation.
But maybe madspin can be impacted.

Now, madspin is indeed configure to have many running executable in idle, and this is typically difficult to track which one is working since the code jumps from one to another depending of the initial/final state of the events that needs to be decayed.

Cheers,

Olivier

Hi Olivier,

sorry, we are still struggling a bit with this and I have to ask again.

In the past (same MG version, only different environment) we have generated ttW with the same setup and it was working fine.
Obviously we are also worried about the validity of the old sample now.

For tW the problem is clear, in the 5FS we can creat tt~ by adding a real emission. We are using the MADSTR plug-in to remove problematic diagrams here.
But is tt~W really equivalent? I don't quite see the same problem? Or is the issue more subtle?

Cheers,
Hannes

Hi,

Here the problem should be the same, you should be able to create
tt~t (or tt~t~) due to the real emission. Or am I wrong and this configuration is not possible?
You should check the real matrix-element to be sure. If you do not hit such configuration then indeed no problem.

Cheers,

Olivier

Hi Olivier,

I don't think you can generate this configuration. You can't generate p p > t t t~.
So I think then we are back to the first question, how to debug the stuck MadSpin process?

Cheers,
Hannes

Ok,

I do not reproduce the issue, so it will be close to impossible to debug.
One would need to add a lot of print statement to know which executable is stalling
and then check the stdin/stdout of that executable to see what the communication issue is.
(or if it is an issue of an infinite loop due to an physical issue)

Cheers,

Olivier

Hi Olivier,

thanks. For what it's worth, the issue only arises for the multiparticle case
define w = w+ w-
generate p p > t t~ w [QCD]

and not for e.g.
generate p p > t t~ w+ [QCD]

And yeah, it appears to depend on the system.

Cheers,
Hannes