Request PSAT in the 1D_COMP restart output - #1593
Conversation
9e08a09 to
5ad0c13
Compare
|
jenkins build this failure_report please |
Did you create those on your own machine? If so, there's a chance that they won't apply to the CI system. I've experienced on a number of occasions that my machine computes result sets that are ever so slightly different from those of the CI system's. Different CPUs, different compilers (or compiler versions), different build flags &c could all be contributing to changes that make locally generated solutions not match those of the CI system. |
Yes. but this one works. Like the jenkins test https://ci.opm-project.org/job/opm-common-PR-builder/10316/ shows. I am okay to undo it and use Jenkins update also. This jenkins request in opm-tests just wants to show that the simulator If you prefer, I can remove the reference update and let jenkins do the work. |
Only until it doesn't. I just merged #1595 generated by the changes in OPM/opm-common#5348 and that has introduced merge conflicts for the
Yes please. I generally prefer Jenkins generating reference solution updates. |
5ad0c13 to
adf30c6
Compare
No compositional regression case asked for the saturation pressure, so the PSAT output path had no coverage: the decks that do request it are the equilibration pair, whose tests are not registered yet, and the black-oil gaslift and spe02 cases never reach the compositional module. The case exercises both branches of the output. The early steps are two-phase, where PSAT is the cell pressure, and the column then gasses out to near-pure CO2 at 150 degC, far above its critical temperature, so the late steps have no saturation pressure and report zero. Between those, a few single-phase steps return a calculated saturation pressure. The reference data is left to the automatic update, which regenerates it from a build carrying the PSAT output.
2 participants
Adds
PSATto theRPTRSTlist of1D_COMP.No compositional regression case asked for the saturation pressure, so the
PSAToutput path had no coverage: the decks that do request it are the equilibration pair, whose tests are not registered yet, and the black-oil gaslift and spe02 cases never reach the compositional module.The case exercises both branches of the output. The early steps are two-phase, where
PSATis the cell pressure, and the column then gasses out to near-pure CO2 at 150 degC, far above its critical temperature, so the late steps have no saturation pressure and report zero. Between those, a few single-phase steps return a calculated saturation pressure, so the solver path is exercised as well as the two trivial ones.The reference data is deliberately not included: it is left to the automatic update, which regenerates it from a build carrying the
PSAToutput. The comparison therefore fails until that lands.Requires the
PSAToutput in OPM/opm-simulators#7389, which in turn requires the solver in OPM/opm-common#5283.