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
We should add solver options (tolerance and max iterations) to the component parameter files. We should also add an option that turns on some diagnostic output, like how many iterations were required for each solve. Currently the parameters are hard coded, but I'm worried they could vary from machine to machine. This would at least allow us to investigate more and tune things. If tolerances are too tight solvers may hit the iteration limit. While not a disaster, it is a big waste of time.
eslickj
changed the title
Next release add Helmholtz EoS solver tolerance and diagnostic options.
Add Helmholtz EoS solver tolerance and diagnostic options.
Apr 4, 2023
We should add solver options (tolerance and max iterations) to the component parameter files. We should also add an option that turns on some diagnostic output, like how many iterations were required for each solve. Currently the parameters are hard coded, but I'm worried they could vary from machine to machine. This would at least allow us to investigate more and tune things. If tolerances are too tight solvers may hit the iteration limit. While not a disaster, it is a big waste of time.
Aside from the saturation density and pressure calculations, all other solves have been migrated to Halley's method from Boost https://www.boost.org/doc/libs/1_80_0/libs/math/doc/html/math_toolkit/roots_deriv.html.
The text was updated successfully, but these errors were encountered: