MadGraph5_aMC@NLO

Poles do not cancel in p p > h t b~ j $$ w+ w- [QCD] (v3.X only)

Asked by Hannes

Hi,

the process
 p p > h t b~ j $$ w+ w- [QCD]
runs fine in MG5_aMC v2.9.9 but not in v3.4.1 (or other MG3 versions as far as I am aware) due to a poles don't cancel error.
I didn't change anything wrt to the MG defaults in each version in both cases. Well we had to change -O3 to -O2 in makefile_loop though (see bug report 1998203).

Is there a way to fix or work around the poles don't cancel issue here?

Probably unrelated but I see these differences as well:
The default coupling orders are different in 3.4.1 and in 2.9.9 (squared vs amplitude level).
And I get this warning in 2.9.9
"Some loop diagrams contributing to this process are discarded because they are not pure (QCD)-perturbation"
But not in 3.4.1.

Cheers,
Hannes

Question information

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

Hi,
in v3, in order to recover the v2 behaviuour for single-top and VBF-like processes (i.e neglecting loop diagrams where you have the W), you need to do
set nlo_mixed_expansion False

This way pole cancelation occurs also in v3 and the run goes on

Best wishes,

Marco

Revision history for this message
Hannes (hannes3) said : 		#2

Hi Marco,

thanks for the quick reply, we will try this.

Cheers,
Hannes

Revision history for this message
Hannes (hannes3) said : 		#3

Hi,

does this also work in 3.3.1 (the version we are currently using in ATLAS)? It doesn't appear to change anything.
Can we replicate the behaviour by using the coupling order restriction?

Cheers,
Hannes

Revision history for this message
marco zaro (marco-zaro) said : 		#4

In principle
set nlo_mixed_expansion False

should work also in 3.3.1, afaik

Best,

Marco

Revision history for this message
Hannes (hannes3) said : 		#5

Hi Marco,

there might be something about the ATLAS setup that makes it not work... what is the effect of the command, is it equivalent to adding the QCD<=N QED<=M restriction or is there more to it?

Cheers,
Hannes

Revision history for this message
Olivier Mattelaer (olivier-mattelaer) said : 		#6

Hi,

I have checked with 3.5.0 (the pre-release of today) compare to 2.9.13 (latest LTS)
and here is the result
2.9.13: INFO: Generated 8 subprocesses with 1072 real emission diagrams, 64 born diagrams and 696 virtual diagrams
3.5.0 (nlo_mixed_expansion True): INFO: Generated 8 subprocesses with 1072 real emission diagrams, 64 born diagrams and 1688 virtual diagrams
3.5.0 (nlo_mixed_expansion False): INFO: Generated 8 subprocesses with 1072 real emission diagrams, 64 born diagrams and 696 virtual diagrams

And I check that both 2.9.13 and 3.5.0 (with set nlo_mixed_expansion False) both have the pole test, soft test and collinear test going trough.

> what is the effect of the command,

The effect of that command is to be able to get back to the behaviour of the 2.x branch where some loop were discarded because they were not pure QCD.

> Is it equivalent to adding the QCD<=N QED<=M restriction or is there more to it?

It is certainly more complex than that, I guess that you can design a loop filter that will also discard those loop, but I do not think that you can achieve that exact effect with the coupling order syntax (at least in general)

Cheers,

Olivier

Can you help with this problem?

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

To post a message you must log in.