GENIE Status Gabriel N. Perdue Simulations for Neutrinos 30 January 2017
Overview • Recent release: GENIE 2.12.2 • Planned releases: - GENIE 3.0 - GENIE 4.0 2 Gabriel Perdue | GENIE news | Simulations for Neutrinos 30 January 2017
GENIE 2.12.2 • Non-physics changes - but important! - Moved z-expansion entries in EGSyst enum to the end of the list so as to preserve the order of (prior) subsequent values. This keeps stored re-weighting values from 2.10.x backwards compatible. • Note, however, that samples generated with 2.12.0 will have problems using stored re-weights with 2.12.2. If you are in this position, we can show you how to patch GENIE to keep it compatible with your weights on disk, but we encourage you to do your next large production sample with the new enum ordering. - Add SingleKaon library to the set loaded by the genie-config script. - Name for TGraph of strange cross sections in ROOT files fixed. 3 Gabriel Perdue | GENIE news | Simulations for Neutrinos 30 January 2017
GENIE 3.0 • The basic idea behind GENIE 3.0 is to choose a set of new default models based on studying agreement with data and considering recent advances on the theory side. We also plan to include technical changes that make changing the physics model simpler (in addition to bug fixes, etc.). - We want to choose a set of models that describe GENIE for the parameter tuning stage to follow. The result of that work will be GENIE 4.0. - We may continue to consider model choices as we begin tuning for 4.0. - We will make comprehensive comparisons to experimental data available as part of the release. - We are currently testing five basic variations. 4 Gabriel Perdue | GENIE news | Simulations for Neutrinos 30 January 2017
GENIE 3.0 - versions • The first two configurations are very close to the current default, with one as close as possible and the other only additionally including the "Empirical" MEC model. • The next update removes the Llewellyn Smith QE model in favor of the Nieves et al model, with RPA and Coulomb effects disabled. It additionally updates the HA FSI model to the most recent version. • The next version represents our "best theory model" - Local Fermi Gas nuclear model - z-expansion for the axial form factor - Nieves et al CCQE with RAP and Coulomb effects - Nieves et al MEC - Berger-Sehgal coherent pion - Berger-Sehgal resonance model - Most recent HA FSI model • We will also have a version with all these changes but using the HN (cascade) FSI model. 5 Gabriel Perdue | GENIE news | Simulations for Neutrinos 30 January 2017
~Current default + MEC (red) GENIE 3.0 ~Current default (black) T [0.2; 0.3 ] GeV T [0.3; 0.4 ] GeV • Evaluating ~400 page plot ∈ ∈ µ µ 1.5 /GeV/n] /GeV/n] book. 2 2 cm cm 1.5 -38 -38 [10 [10 1 • Lots of effort currently being µ µ T T ∂ ∂ / / 1 µ µ θ θ Cos Cos invested in understanding ∂ ∂ / 0.5 / σ σ 2 2 ∂ ∂ 0.5 signal definitions, normalization, etc. 0 0 − 1 − 0.5 0 0.5 1 − 1 − 0.5 0 0.5 1 Cos Cos θ θ miniboone_nuccqe_2010 miniboone_nuccqe_2010 • Some datasets pull the µ µ trunk:G00_00a:miniboone_fhc trunk:G00_00a:miniboone_fhc trunk:G00_00b:miniboone_fhc trunk:G00_00b:miniboone_fhc trunk:G16_01a:miniboone_fhc trunk:G16_01a:miniboone_fhc trunk:G16_02b:miniboone_fhc trunk:G16_02b:miniboone_fhc choices in one direction, other T [0.4; 0.5 ] GeV T [0.5; 0.6 ] GeV ∈ ∈ µ µ datasets pull in another. /GeV/n] /GeV/n] 3 MiniBooNE Best 2 2 2 cm cm • We will make the comparison CCQE -38 -38 theory, [10 [10 2 1.5 µ µ T T plots available with the ∂ ∂ / / hN FSI µ µ θ θ Cos Cos 1 ∂ ∂ release so experiments have / (blue) / σ σ 1 2 2 ∂ ∂ 0.5 a sense for the impact of 0 0 − 1 − 0.5 0 0.5 1 − 1 − 0.5 0 0.5 1 changes. Cos Cos θ θ miniboone_nuccqe_2010 miniboone_nuccqe_2010 µ µ trunk:G00_00a:miniboone_fhc trunk:G00_00a:miniboone_fhc trunk:G00_00b:miniboone_fhc trunk:G00_00b:miniboone_fhc trunk:G16_01a:miniboone_fhc trunk:G16_01a:miniboone_fhc trunk:G16_02b:miniboone_fhc trunk:G16_02b:miniboone_fhc 6 Gabriel Perdue | GENIE news | Simulations for Neutrinos Simple changes (green) 30 January 2017
GENIE 4.0 - GENIE + Professor Energy Frontier Slide by C. Andreopoulos 7 Gabriel Perdue | GENIE news | Simulations for Neutrinos 30 January 2017
GENIE 4.0 Fundamentally, the idea of Professor is to: ● Reduce the exponentially expensive process of brute-force tuning to a scaling closer to a power law in the number of parameters. ● Allow for massive parallelisation and systematically improve the scan results by use of a deterministic parameterisation of the generator’s response to changes in the steering parameters. [H.Schulz] Slide by C. Andreopoulos 8 Gabriel Perdue | GENIE news | Simulations for Neutrinos 30 January 2017
GENIE 4.0 GENIE / Professor interface GENIE comparison 1 GENIE comparison 2 GENIE comparison 3 ... description description description description description description description description description description description Bin #10 Bin #11 Bin #1 Bin #2 Bin #3 Bin #4 Bin #5 Bin #6 Bin #7 Bin #8 Bin #9 Pass to Professor: ● A single 1-D array with all data ● A single 1-D array with all predictions (per given model configuration and parameter values) ● A single 2-D covariance matrix Slide by C. Andreopoulos 9 Gabriel Perdue | GENIE news | Simulations for Neutrinos 30 January 2017
GENIE 4.0 Current status Borrowed slides are bit old here... Batch system readiness ● Main MC production site at Lyon now functional / scripts available for that site ○ Several production tasks carried out at Lyon ■ Processing for v3.0.0 characterization and for Professor/GENIE tuning exercise ● An additional substantial resource at RAL/PPD Tier2 also functional / available as a backup for more CPU-intensive stages of the Professor/GENIE tuning exercise Code development ● Professor able to produce UserPhysicsOptions.xml steering cards - Done ● Automated GENIE MC production scripts - Completed and tested ● Scripts for automated configuration of GENIE Comparisons - on going ● Code for linearizing data and MC - Almost completed for MinibooNE and T2K comparisons on-going ● Finalizing output data format (data, predictions, covariances) for Professor - Starting this week Next steps ● Complete migration of integrated cross-section (bubble chamber) and ...lots of progress hadronization comparisons in the new framework completed here ○ Compatibility with the new system / scripts ○ Add \chisquare calculations for quantitative model characterizations, as done for MB/T2K ○ Linearise data and predictions to append to standardised GENIE output for Professor and include in tunes Slide by C. Andreopoulos 10 Gabriel Perdue | GENIE news | Simulations for Neutrinos 30 January 2017
Schedule • GENIE 3.0 "soon" - finalizing comparisons - deciding on technical changes • GENIE 4.0 is on an aggressive schedule also. - the goal is this calendar year - lots of technical progress already at Liverpool - but many thorny physics issues (dataset tensions, experiment- theory tensions, opinions about model selection, etc.) to sort out... 11 Gabriel Perdue | GENIE news | Simulations for Neutrinos 30 January 2017
Back-up 12 Gabriel Perdue | GENIE news | Simulations for Neutrinos 30 January 2017
New models in GENIE 2.12 • Valencia MEC • Local Fermi Gas (nuclear model) • Nieves et al CCQE model • Oset modifications to cascade (hN) FSI model • Kaon FSI • z-expansion form of the Axial Form Factor for QE • QEL hyperon production • Berger-Sehgal coherent pion model • Updated Rein diffractive pion model • Energy-dependent MA model (Kuzmin-Naumov) 13 Gabriel Perdue | GENIE news | Simulations for Neutrinos 30 January 2017
Recommend
More recommend
