MadGraph5_aMC@NLO

trouble at the end of the genenrating

Asked by Meng Lu

i met some troubles when i run madgraph at fermi's server. fermi's account has 24 core, and all work ok but problems occur when the generating is going to finish

INFO: Idle: 0, Running: 24, Completed: 179 [ 1h 38m ]

INFO: Idle: 0, Running: 5, Completed: 187 [ 1h 44m ]
INFO: Idle: 0, Running: 4, Completed: 188 [ 1h 45m ]
INFO: Idle: 0, Running: 3, Completed: 189 [ 1h 47m ]
INFO: Idle: 0, Running: 2, Completed: 190 [ 2h 10m ]

INFO: Idle: 0, Running: 1, Completed: 191 [ 2h 53m ]

and it took a very long time to complete the job or just stop there until it was kill by the daemon.

could someone told what has cause this and how to solve it.

thanks in advance

Question information

Language:
English Edit question
Status:
Solved
For:
MadGraph5_aMC@NLO Edit question
Assignee:
No assignee Edit question
Solved by:
Olivier Mattelaer
Solved:
Last query:
Last reply:
Revision history for this message
Olivier Mattelaer (olivier-mattelaer) said : 		#1

It means that one integral is much more unstable than all the other one.
This maybe related to too soft cut or to other physical reason.

Cheers,

Olivier

> On Oct 13, 2016, at 06:23, kojo <email address hidden> wrote:
>
> New question #402977 on MadGraph5_aMC@NLO:
> https://answers.launchpad.net/mg5amcnlo/+question/402977
>
> i met some troubles when i run madgraph at fermi's server. fermi's account has 24 core, and all work ok but problems occur when the generating is going to finish
>
> INFO: Idle: 0, Running: 24, Completed: 179 [ 1h 38m ]
>
>
> INFO: Idle: 0, Running: 5, Completed: 187 [ 1h 44m ]
> INFO: Idle: 0, Running: 4, Completed: 188 [ 1h 45m ]
> INFO: Idle: 0, Running: 3, Completed: 189 [ 1h 47m ]
> INFO: Idle: 0, Running: 2, Completed: 190 [ 2h 10m ]
>
> INFO: Idle: 0, Running: 1, Completed: 191 [ 2h 53m ]
>
>
> and it took a very long time to complete the job or just stop there until it was kill by the daemon.
>
> could someone told what has cause this and how to solve it.
>
> thanks in advance
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Revision history for this message
Meng Lu (meng-lu) said : 		#2

thanks, Oliver,

 but it is the same condition as i run the process in my own laptop and it always occurs. and i wonder the label "running" is the cores at running or not? it could take more than 2 hours to let the "running" number change from 2 to 1, is it possible to spend so long time to do a integral?

Revision history for this message
Best Olivier Mattelaer (olivier-mattelaer) said : 		#3

> and i wonder the label "running" is the cores at
> running or not?

not always but most of the time yes.
(Some times it is on running waiting for IO or for control status —mainly if something went wrong—)

> is it possible to spend so long time to do a
> integral?

Yes it is possible. It can sometimes takes much longer than that.
Everything depends of the complexity of your matrix-element (number of amplitude/number of matrix-element/presence of loop/typical stability of the loop/…) and of your cuts.

Cheers,

Olivier

> On Oct 13, 2016, at 12:45, kojo <email address hidden> wrote:
>
> Question #402977 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/402977
>
> kojo posted a new comment:
> thanks, Oliver,
>
> but it is the same condition as i run the process in my own laptop and
> it always occurs. and i wonder the label "running" is the cores at
> running or not? it could take more than 2 hours to let the "running"
> number change from 2 to 1, is it possible to spend so long time to do a
> integral?
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Revision history for this message
Meng Lu (meng-lu) said : 		#4

Olivier,

thank you very much.

Revision history for this message
Meng Lu (meng-lu) said : 		#5

Thanks Olivier Mattelaer, that solved my question.