Added SBN (shared) GENIE configuration for BNB and NuMI #492
Conversation
|
Thanks, @PetrilloAtWork! I agree SBND should also use these common settings. I'll take a look at your settings soon so we have this PR merged quickly, but it may take a bit longer for SBND to use it. SBND should review what changes this brings us, especially waht you noted about the beam intensity, before we can adopt it. Also added @miquelnebot as SBND sim convener. |
|
Hi @PetrilloAtWork, I took a deeper look and compared with the current settings uded in SBND (using My only question is about the intensity change you mention in the description. We are already using 5e12 POT per spill, which seems the same as in this PR. Or are you referrerring to something else?
|
|
In 2024, intensity barely went above 4·10¹² PoT/spill, while we insist on the "MicroBooNE" value of 5·10¹² PoT/spill. In 2023, it was even less. Hence my point. |
|
@marcodeltutto If you are satisfied with @PetrilloAtWork answer, can you approve the PR? if not, what else would you like clarification on or changes to this? |
|
@ibsafa I had a few more questions above but I prefer to get this going since those are minor things. Approving now. |
| # | ||
| beam_structure_BNB_2024: { | ||
|
|
||
| # enable bunched structure |
There was a problem hiding this comment.
I don't think this adds the bunched structure, but rather the neutrino time from the flux file, so from the proton-on-target to neutrino production, plus from the production point to the interaction vertex. Maybe it should say "adds in the neutrino time from proton-on-target to interaction point (if available in flux file)"
| ### neutrino flux parameters | ||
| ### | ||
|
|
||
| DetectorLocation: "MINOS-NearDet" # location name for flux window |
There was a problem hiding this comment.
Shall we change this or will it break things? :D
| # data-like samples, but for the far detector this is a very inefficient | ||
| # generation; here we disable it. | ||
| # | ||
| EventsPerSpill: @erase |
There was a problem hiding this comment.
Is @erase == 0? From the docs, POTPerSpill is used if EventsPerSpill == 0
Description
This request stores in
sbncodea configuration template forGENIEGenmodule that include common settings like the time structure of the beam, its intensity, ν-Ar cross sections, and a few sparse others. Conversely, flux and simulation volumes are intended to be specified by the derivative configurations (although the latter does have a default volume specified).Both BNB and NuMI beams are described.
This PR is a dependency of SBNSoftware/icaruscode#785: ICARUS is going to use this PR as base configuration.
Choice of parameters
The configuration almost completely replicates the one from SBND (
genie_sbnd.fcl), although the difference with the current ICARUS is almost completely insignificant beside the enabling of beam time structure, which is the issue that sparked this PR.For the beam timing, SBND figures were used; the bunch period matches the one measured by ICARUS, and the width of the bunch is... well, probably too narrow, but I don't have a "naked" measure of it (even MicroBooNE's folds detector effects into it).
For the beam intensity, I stuck to the "official" values used in the SBN proposal, which were reasonably close to what MicroBooNE had seen but that are hopelessly higher than what SBN detector have observed, typically by a factor ~1.5. Nevertheless, changing these figures needs to be discussed by the experiments and especially by SBND, which is the most affected (as long as events include only one interaction, intensity changes can be approximated by just scaling the total exposure; but SBND does see interaction pile up).
Details
The new configuration fragments are all in a single
genie_beam_settings.fcl.It is structured in a timing section where the beam time structure is described, followed by a section with templates for
GENIEGenconfiguration. Each of the section is split between BNB and NuMI.Also, some beam time structures configuration fragments there are meant to be the default/current setting (e.g.
beam_parameters_BNB), and others to be immutable snapshots of the state at a certain time (e.g.beam_parameters_BNB_2024).You may thing of
beam_parameters_NuMI_2024as an internal, sad sad joke.Review
I summon for review:
I invite the reviewers to add more reviewers as they feel fit.
I tested the BNB and NuMI configurations in ICARUS; I did not attempt to modify SBND configurations to use these configuration templates.
I would like the reviewers to also express about the parameters I removed compared to SBND configuration, which I believe I have identified as unused (and seriously misleading; one example:
BeamCenter: [ 350, 1400, 0 ]).I would also like SBND to consider to use this configuration as base for its GENIE jobs.
One presentation about this change was given to ICARUS event selection working group (SBN DocDB 39065). I think I'll give a similar one in a SBN analysis infrastructure meeting (same document number), but I am not sure when the next one will be called — and I would prefer not to wait for that.