Repository navigation
Conversation
bd713
added
ci: run CUDA builds
ci: run integrated tests
| GEOS_LOG_LEVEL_RANK_0( logInfo::TimeStep, GEOS_FMT( "New dt = {}", stepDt ) ); | ||
|
|
||
| // notify the solver statistics counter that this is a time step cut | ||
| getIterationStats().updateTimeStepCut(); |
There was a problem hiding this comment.
| getIterationStats().writeIterationStatsToTable(); |
Is it intentional not to save the iteration stats for the CoupledSolver?
There was a problem hiding this comment.
Not intentional, thanks!
Modification applied.
| m_convergenceData.addRow( residualsNormCells ); | ||
|
|
||
| if( !m_CSVOutputOpened ) | ||
| if( MpiWrapper::commRank() == 0 ) |
There was a problem hiding this comment.
Should it apply to the whole writeConvergenceStatsToTable() method?
| real64 const setupTime = m_setupTime; | ||
| real64 const solveTime = m_solveTime; | ||
| resetSolverLinearTime(); | ||
|
|
There was a problem hiding this comment.
Should it go before the first if? Maybe a scoped if can help.
There was a problem hiding this comment.
I would keep the order.
The first reset must only happen once a time step has converged, otherwise the time spent on a cut first step will be lost.
| m_logStream << m_iterationCSVFormatter->headerToString( ); | ||
| m_CSVOutputOpened = true; | ||
| m_CSVOutputOpened = true; |
There was a problem hiding this comment.
A bit of renaming could clarify: m_logStream -> m_csvStream
| // so writeIterationStatsToTable() early-returns: no rows expected. | ||
| } | ||
|
|
||
| TEST( CoupledSolverStatisticsCSVTest, checkCSVOutput ) |
There was a problem hiding this comment.
IIUC, as this is the first unit-test that tests **Poromechanics solvers, should this test file be renamed testPoromechanics.cpp? (something like that)
To be more general, regrouping high-level integration-tests, can give an opportunity to optimize the loading time.
There was a problem hiding this comment.
I kept the test with the feature it checks rather than the physics it runs.
But I moved it next to the testSolverStats.
All its assertions are about the iteration/convergence CSV output, not the physics.
Uh oh!
There was an error while loading. Please reload this page.