You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Sensitivity files for Serpent 2.1.31 include (by default or with additional options) data from intermediate latent generations. The naming of these variables overwrites the "correct" data, e.g. data for sensitivity of keff is taken from the second to last latent generation rather than all latent generations. This issue was reported on the discussion forum
doesn't contain logic for these N_GEN names, and then overwrites the useful information, e.g. data from ADJ_PERT_KEFF_SENS_9_GEN is stored in place of ADJ_PERT_KEFF_SENS as sensitivities["keff"]
Expected outcome
Store data from all latent generations, ADJ_PERT_KEFF_SENS as the actual data, with options to store data from latent generations as well.
Related to GH issue CORE-GATECH-GROUP#365, where the use of the new Serpent
sens opt history 1 option prints out additional sensitivity
data. In this mode, sensitivities using 1<=n<N latent generations
are printed, where N is the number specified using sens opt latgen
argument. This creates new arrays of the form
ADJ_PERT_SENS_<RESP>_<N>_GEN
ADJ_PERT_SENS_<RESP>_<N>_GEN_E_INT
to the sensitivity files. As reported in the issue,
serpentTools would store the last set of <N>_GEN as the actual
response, so the value stored would reflect data using one fewer
latent generations.
The sensitivity test file is modified to include four new arrays,
ADJ_PERT_keff_SENS_1_GEN
ADJ_PERT_keff_SENS_1_GEN_E_INT
ADJ_PERT_fis2flx_SENS_1_GEN
ADJ_PERT_fis2flx_SENS_1_GEN_E_INT
These arrays are identical the non _1_GEN values, except the
very first value and associated uncertainty are set to zero.
Running the sensitivity file test will fail due to this
change, as the reference data have not been changed in the file,
but are overwritten as reported in the issue.
Summary of issue
Sensitivity files for Serpent 2.1.31 include (by default or with additional options) data from intermediate latent generations. The naming of these variables overwrites the "correct" data, e.g. data for sensitivity of keff is taken from the second to last latent generation rather than all latent generations. This issue was reported on the discussion forum
Code for reproducing the issue
New
KEFF_SENS
data are provided asActual outcome including console output and error traceback if applicable
The method for processing these arrays
serpent-tools/serpentTools/parsers/sensitivity.py
Lines 231 to 236 in b1aa255
doesn't contain logic for these
N_GEN
names, and then overwrites the useful information, e.g. data fromADJ_PERT_KEFF_SENS_9_GEN
is stored in place ofADJ_PERT_KEFF_SENS
assensitivities["keff"]
Expected outcome
Store data from all latent generations,
ADJ_PERT_KEFF_SENS
as the actual data, with options to store data from latent generations as well.Versions
Please provide the following:
Per the reporting user
serpentTools.__version__
0.9.0
python --version
N/Agit describe
The text was updated successfully, but these errors were encountered: