MadGraph5_aMC@NLO

Rerunning single failed job for NLO gridpack production

Asked by LouisHelary

Dear all,

I'm trying to generate a gridpack for the following process:
Defined multiparticle p = g u c d s u~ c~ d~ s~
Defined multiparticle j = g u c d s u~ c~ d~ s~
Defined multiparticle l+ = e+ mu+
Defined multiparticle l- = e- mu-
Defined multiparticle vl = ve vm vt
Defined multiparticle vl~ = ve~ vm~ vt~
define p = g u c d s b u~ c~ d~ s~ b~
define j = g u c d s b u~ c~ d~ s~ b~
define e = e+ e-
define v = vl vl~
generate p p > e+ e- a [QCD]
generate p p > e+ e- a j [QCD]
generate p p > e+ e- a j j [QCD]

I have tried to run the gridpack creation on several different types of machine (single multi-core machine, or on LSF and SGE clusters). The problem is that the number of subjobs that needs to be run is very large ( Idle: 3416, Running: 16, Completed: 0 [ current time: 20h10 ] ), and some of these subjobs can take up to five days to finish. I have therefore never managed to run in one go the gridpack creation, as I always ended-up with a few of these jobs getting killed on the LSF or on the SGE cluster, and unfortunately a powercut killed my calculation on the single multi-core machine after 3 weeks of calculations...

So I would like to understand if it is possible to rerun a few failed PROCNLO_loop_sm_0/SubProcesses/P0_XX/GFYY directories by hand, and what would be the procedure to then finish the gridpack creation?

I'm trying to generate events with MadGraph5_aMC@NLO v2_3_3.

Thanks in advance for your help.

Best,

Louis

Question information

Language:
English Edit question
Status:
Solved
For:
MadGraph5_aMC@NLO Edit question
Assignee:
marco zaro Edit question
Solved by:
LouisHelary
Solved:
Last query:
Last reply:
Revision history for this message
marco zaro (marco-zaro) said : 		#1

Hi Louis,
the process you are trying is a quite complicated one. are you setting the photon-isolation cuts in the run_card and the iso_em variable to true?
In any case, I would suggest you to try these simpler solutions:
1) z-photon instead of e+ e- photon
2) stop at 1 extra jet for the merging

Let me know,

Cheers,

Marco

Revision history for this message
LouisHelary (louishelary) said : 		#2

Hi Marco,

Thanks for your answer. When you say that the computation is challenging, does that mean that it is impossible? Even if I would manage to get it over, the results would not be to trust?

To answer your questions yes I'm setting isolation with the following parameters:

10 = ptgmin ! Min photon transverse momentum
 -1.0 = etagamma ! Max photon abs(pseudo-rap)
 0.3 = R0gamma ! Radius of isolation code
 2.0 = xn ! n parameter of eq.(3.4) in hep-ph/9801442
 0.025 = epsgamma ! epsilon_gamma parameter of eq.(3.4) in hep-ph/9801442
 True = isoEM ! isolate photons from EM energy (photons and leptons)

And I already managed to get some results with 0,1j at NLO, but I'm interested by the two jet topology and specifically by the kinematic differences between these jets so I would prefer to push into getting 0,1,2j topology if the computation is of course possible.

So my questions remain, is there a way to recover a single failed subjobs with NLO grid pack generation? And if yes what would be the procedure?

Again thanks for your answer,

Best,

Louis

Revision history for this message
marco zaro (marco-zaro) said : 		#3

Hi Louis,
it is not impossible, but it may take very long time to run.
This is why I would suggest to try to use Z gamma and decay the Z afterwards with MadSpin.
Finally, there is no way to re-run a single job, unfortunately.

Cheers,

Marco

Revision history for this message
LouisHelary (louishelary) said : 		#4

Hi Marco,

 Thanks for your answer Ill try the Zgamma+ mad spin approach then.

Best,

Louis