Changing number of jobs

Asked by Artem Golovatiuk

Hi,

I have tried to find something about it, but not successfully.
I am trying to run MG on my University's cluster, and the question is if there is a possibility to increase number of jobs by hands.

As far as I understand, jobs are parts of phase-space integration, so it seems that there should not be any restriction from above on their number. And it would help me to accelerate the calculations, for example, if I take 16 CPUs from cluster, but I get 4 or 8 jobs by default in some process.
There certainly should be a way to optimize this process, but I did not manage to find it.

Thanks

Question information

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

Hi,

All the parameter that we have available for that are explained in the appendix of this paper
arXiv:1507.00020<http://arXiv.org/abs/arXiv:1507.00020> (those parameters have been introduced for loop-induced but some of them are available for normal LO computaiton —as stated in the paper--)

Note that having more jobs than the nominal one, required then a huge communication between the job. (In this context it means a lot of file written on disk)
And therefore if you have only 4-8 job it is likely that increasing to 16 will actually not help at all.

Cheers,

Olivier

Note that such options are only available for tree
On 8 Apr 2017, at 19:32, Artem Golovatiuk <<email address hidden><mailto:<email address hidden>>> wrote:

New question #615607 on MadGraph5_aMC@NLO:
https://answers.launchpad.net/mg5amcnlo/+question/615607

Hi,

I have tried to find something about it, but not successfully.
I am trying to run MG on my University's cluster, and the question is if there is a possibility to increase number of jobs by hands.

As far as I understand, jobs are parts of phase-space integration, so it seems that there should not be any restriction from above on their number. And it would help me to accelerate the calculations, for example, if I take 16 CPUs from cluster, but I get 4 or 8 jobs by default in some process.
There certainly should be a way to optimize this process, but I did not manage to find it.

Thanks

--
You received this question notification because you are an answer
contact for MadGraph5_aMC@NLO.

Can you help with this problem?

Provide an answer of your own, or ask Artem Golovatiuk for more information if necessary.

To post a message you must log in.