When running sims with conservation of chemical potential, the computational cost increases due to these projections in the post processing when computing surface fluxes and exporting the retention to XDMF:
A simple test checking the type of the temperature should be sufficient to greatly optimise the computational time in some cases.
Even better, it'd be super to sort out how to make it work with the expression type....
When running sims with conservation of chemical potential, the computational cost increases due to these projections in the post processing when computing surface fluxes and exporting the retention to XDMF:
https://github.com/RemDelaporteMathurin/FESTIM/blob/41f11c588366d0555e0f3c3e712e67c10d88bab2/FESTIM/post_processing.py#L80-L82
and
https://github.com/RemDelaporteMathurin/FESTIM/blob/41f11c588366d0555e0f3c3e712e67c10d88bab2/FESTIM/post_processing.py#L411-L413
The latter could be avoided when the type of the temperature is not
"expression"
. See the MWE below:A simple test checking the type of the temperature should be sufficient to greatly optimise the computational time in some cases. Even better, it'd be super to sort out how to make it work with the expression type....