No shower_card.dat when running ```./scripts/setup_process_NLO.sh ggF-SMEFTatNLO```

[maxgalli]

After setting up the environment and running cp setup/SMEFTatNLO__restrict_ggFshape.dat MG5_aMC_v2_6_7/models/SMEFTatNLO/restrict_ggFshape.dat in order to have the needed restrict card in the proper place, I tried to run ./scripts/setup_process_NLO.sh ggF-SMEFTatNLO, but this doesn't seem to produce any shower_card.dat or shower_card_default.dat inside MG5_aMC_v2_6_7/ggF-SMEFTatNLO. This then causes a failure when preparing the gridpacks.

For simplicity, I reduced the card to 0 jet only:

import model SMEFTatNLO-ggFshape

generate g g > H QED=1 NP=0 [noborn=QCD]

output ggF-SMEFTatNLO

but this seems to happen also with the card in master.

Am I doing something wrong?

Thanks!

[ajgilbert]

Hi, so I think the issue is these loop induced processes are treated, a bit counter-intuitively, as LO processes in terms of running MG. So if you use the normal LO workflow in EFT2Obs, setup_process.sh etc, it should work. I actually hope to have a few small updates & some instructions for getting the correct ggH prediction soon - I found a few subtleties in how we have to combine the pure loop, pure tree and tree-loop interference contributions.

[maxgalli]

Ok, thanks. Just for my own understanding (and to throw in some interesting material): the procedure followed here is somehow similar to the one suggested at page 20 here? Meaning that we first generate the baseline sample and then compute at NLO using reweighting?

In principle this card

import model SMEFTatNLO-ggFshape

generate g g > H QED=1 QCD=2 NP=2 [QCD]

output ggF-SMEFTatNLO-mod

works but it's probably more computationally expensive and it seems to fail at another point of the process that I still have to investigate.

[ajgilbert]

Right, it's a similar idea, except here we start from the loop-level SM process:

generate p p > H QED=1 NP=0 [noborn=QCD]

then reweight to either:

• Loop only: change process p p > h NP=2 QED=1 [noborn=QCD]
• Tree only change process p p > h NP=2 QED=1
• Loop-tree interference change process p p > h NP=2 QED=1 [virt=QCD]

[ajgilbert]

Some notes here, will add more over the next week or so:

https://github.com/ajgilbert/EFT2Obs/blob/ggH-example/docs/HiggsGluonFusion.md