Add genfragments for PbPb 5.36 TeV Madgraph production

[stahlleiton]

This PR adds the genfragments for official PbPb 5.36 TeV (Run3) production of Madgraph generator in the directory: genfragments/PbPb_5p36TeV/MadGraph5_aMCatNLO .

The genfragments are created using as templates the central production Powheg genfragments for Run3Summer23 defined in https://github.com/cms-PdmV/GridpackFiles/tree/master/Cards/MadGraph5_aMCatNLO .

The main differences in the process parameters compared to 2018 (Run 2) were:

• TT: Kept the paramaters: "TimeShower:mMaxGamma = 1.0" to be consistent with the POWHEG gen fragment (the central gridpacks use TimeShower:mMaxGamma = 4.0 for MadGraph while 1.0 is used for Powheg so seems inconsistent).
• TT: Is the parameter "TimeShower:MEcorrections = on" still needed? (kept it from 2018 but I see it is not used in 2023)
• DY: Added the parameter "BeamRemnants:primordialKThard=2.48" (was not in 2018 but appears in 2023 central gridpacks).

@sarteagae @DickyChant @menglu21 @bbilin

[sihyunjeon]

1 and 4 gev difference was due to the ignorance of the impact it would cause (and it should be negligible actually)
we gave 4 mmaxgamma consistently (even for powheg) in 23 samples as our ME level mll is 4gev minimum unless specified

[sihyunjeon]

me corrections came from historic decisions way before my times so i don't know why we had it in the first place but from discussions in https://gitlab.com/Pythia8/releases/-/issues/49 we decided to ignore

[sihyunjeon]

dy pt tune value is from intrinsic kt tune GEN-22-001(?)

[sihyunjeon]

one thing also small but minor to comment in one of your previous commits was that you don't tune the SM parameters which are slightly different from default sm ufo files in madgraph, if you care about 100% synchronization

https://github.com/cms-PdmV/GridpackFiles/tree/master/Campaigns/Run3Summer22/MadGraph5_aMCatNLO/ModelParams

[stahlleiton]

1 and 4 gev difference was due to the ignorance of the impact it would cause (and it should be negligible actually) we gave 4 mmaxgamma consistently (even for powheg) in 23 samples as our ME level mll is 4gev minimum unless specified

ok, so if I understand correctly, now for both Powheg and Madgraph, I should set the mmaxgamma paramater to 4 in the TTbar gen fragments, right?

[stahlleiton]

me corrections came from historic decisions way before my times so i don't know why we had it in the first place but from discussions in https://gitlab.com/Pythia8/releases/-/issues/49 we decided to ignore

ok, then I will remove the parameter "TimeShower:MEcorrections" to keep it consistent with pp central if not relevant anymore

[stahlleiton]

@sihyunjeon : just updated the genfragment: https://github.com/cms-sw/genproductions/pull/3684/files#diff-5376b9df1a6173eed869d93f4cdca675efd1e26cfbbc10acdbeb75a10816ebd3
Let me know if you have further comments on the MadGraph genfragments, thanks

[DickyChant]

Let me leverage this thread:

@stahlleiton and I feel that it would be nice to have integration from PdmV side as in the common background handling procedure for those HIN gridpacks and even requests. @sihyunjeon @agrohsje and @Cvico I think I approached in the past but let me ask again is it simple for you guys to handle this maybe?

[sihyunjeon]

@stahlleiton For my education, i am not fully aware how the UE tunes were set for HIN, do we just take 13.6 TeV parameters and cross sections?

[DickyChant]

@stahlleiton For my education, i am not fully aware how the UE tunes were set for HIN, do we just take 13.6 TeV parameters and cross sections?

I think this is what was doing in the past (i.e inheriting CP5 from pp for 5.02 TeV) and for HI there is special mixing from HI dedicated Minbias sample

[sihyunjeon]

@DickyChant we can take those over easily, up to the analyzers and HIN L2 + mc contacts. we did ignore BPH or HIN groups from the beginning since workers (myself and alex) were not part of those groups in the past and was relatively less sure about MC business in those groups. the cores we have are also relatively free as well for now since majority of the samples we had are almost there

[stahlleiton]

@stahlleiton For my education, i am not fully aware how the UE tunes were set for HIN, do we just take 13.6 TeV parameters and cross sections?

How feasible would be to generate a tune for 5.36 TeV?

To answer your question, yes, during Run 2, we simply used the same tune as used in central pp production (CP5 tuned on 13 TeV)

[DickyChant]

@stahlleiton For my education, i am not fully aware how the UE tunes were set for HIN, do we just take 13.6 TeV parameters and cross sections?

How feasible would be to generate a tune for 5.36 TeV?

To answer your question, yes, during Run 2, we simply used the same tune as used in central pp production (CP5 tuned on 13 TeV)

New tune we need UE measurements e.g. hadron spectre

[DickyChant]

One thing also just come to my mind:

don’t we need to adjust kT based on our nice GEN PRL scaling? @sihyunjeon @agrohsje @mseidel42

[mseidel42]

One thing also just come to my mind:

don’t we need to adjust kT based on our nice GEN PRL scaling? @sihyunjeon @agrohsje @mseidel42

Ah yes, would actually be good to use the scaling 👍

[stahlleiton]

@DickyChant we can take those over easily, up to the analyzers and HIN L2 + mc contacts. we did ignore BPH or HIN groups from the beginning since workers (myself and alex) were not part of those groups in the past and was relatively less sure about MC business in those groups. the cores we have are also relatively free as well for now since majority of the samples we had are almost there

If it helps, here is a PR adding the HIN campaign : cms-PdmV/GridpackFiles#151

[DickyChant]

One thing also just come to my mind:
don’t we need to adjust kT based on our nice GEN PRL scaling? @sihyunjeon @agrohsje @mseidel42

Ah yes, would actually be good to use the scaling 👍

I got the slope from GEN-22-001 and did the calculation by hand.

>>> log10_new_value = math.log10(2.48) + (math.log10(5.36) - math.log10(13.6)) * 0.162
>>> 10 ** log10_new_value
2.1327653364672026
>>> 

[stahlleiton]

One thing also just come to my mind:
don’t we need to adjust kT based on our nice GEN PRL scaling? @sihyunjeon @agrohsje @mseidel42

Ah yes, would actually be good to use the scaling 👍

I got the slope from GEN-22-001 and did the calculation by hand.

>>> log10_new_value = math.log10(2.48) + (math.log10(5.36) - math.log10(13.6)) * 0.162
>>> 10 ** log10_new_value
2.1327653364672026
>>> 

thanks. Updated the value in 805b7ab

[stahlleiton]

@sihyunjeon , @bbilin , @menglu21 : do you have further comments on the MadGraph gen fragments for PbPb? If not, could you proceed to merge this PR? Thanks

[stahlleiton]

one thing also small but minor to comment in one of your previous commits was that you don't tune the SM parameters which are slightly different from default sm ufo files in madgraph, if you care about 100% synchronization

https://github.com/cms-PdmV/GridpackFiles/tree/master/Campaigns/Run3Summer22/MadGraph5_aMCatNLO/ModelParams

Here is the PR #3688 fully synchronizing the SM parameters