There is a strange interaction when using MPI and OpenMP on some
hardware/MPI implementations. I a serial run omp_get_num_procs() would
return the number of processors but when started under mpirun it would
always return 1.
With this we now allow users to use any amount of threads.
While we reported that we used the number of threads that were passed
on the command line, we never really used it for OpenMP but always
sticked to two unless environment variable OMP_NUM_THREADS was set.
Note that because the ThreadManager in opm-models would always use the
command line option and hence the linearizer would use that number of
threads.
Please note that the only use of OpenMP in opm-common (volume
calculation in EclipseGrid) is not effected by this as it happens
before we set the number of OpenMP threads.
Following commits f4f8c033d and c7016854d (PR #4286), the "vanguard"
no longer maintains an internal Deck data member. Don't pretend
that it does by providing member functions to access that data
member. If someone tries to actually call these member functions
then they will get an unfriendly diagnostic message and a build
failure.
this is a struct that holds the information transferred
from the simulator prior to vanguard instantation.
this way we can avoid the static members for state,
which causes issues for serialization (in particular, it makes
it impossible to compare two instances like we do in tests).
This commit renames the previously introduced command line option
ExtraConvergenceOutput (--extra-convergence-output) into the more
descriptive
OutputExtraConvergenceInfo (--output-extra-convergence-info)
Suggested by: [at]OPMUSER
This commit enables outputting non-linear convergence metrics, i.e.,
the MB and CNV values, per phase, for each non-linear iteration in
each timestep. If the user passes the option value "iterations" to
the --extra-convergence-output command line option, this commit will
create a new output file, CASE.INFOITER, that holds
* report step
* time step within that report step
* elapsed time
* MB and CNV values per phase
* well convergence status
for each non-linear iteration.
We use an asynchronous file writing procedure and confer ownership
of the report step's unprocessed convergence reports to this
procedure just before the end of
SimulatorFullyImplicitBlackoilEbos::runStep()
At that point, the convergence reports are about to go out of scope.
The asynchronous protocol uses a dedicated queue of output requests,
class ConvergenceReportQueue, into which the producer-i.e., member
function runStep()-inserts new convergence reports and from which
the output thread, ConvergenceOutputThread::writeASynchronous(),
retrieves those requests before writing the file data.
This commit introduces a new helper class,
ConvergenceOutputConfiguration
which parses comma separated option strings into a runtime
configuration object for whether to output additional convergence
information and, if so, what information to output.
Supported option string values are
* "none" -- Dont want any additional convergence output.
* "steps" -- Want additional convergence output pertaining to the
converged solution at the end of each timestep.
* "iterations" -- Want additional convergence output pertaining to each
non-linar ieration in each timestep.
Option value "none" overrides all other options. In other words, if the
user requests "none", then there will be no additional convergence
output, even if there are other options in the option string.
We add a new option, ExtraConvergenceOutput (command line option
--extra-convergence-output), which takes a string argument expected
to be a comma separated combination of these options. The default
value is "none". Finally, make the INFOSTEP file output conditional
on the user supplying "steps" as an argument to the new option.
This enables outputting a formatted record of the limiting MB and
CNV quantities as time and non-linear iterations progress. This, in
turn, is intended for diagnostic and analysis purposes and will not
be output unless specifically requested.
In particular, add a new type,
ConvergenceReport::ReservoirConvergenceMetric
which captures the convergence metric type (MB or CNV) along with
the associate phase and numerical value of the convergence metric.
We add a vector of these convergence metric objects as a new data
member of the ConvergenceReport.
Finally, foreshadowing the intended use case, also store the report
time in the ConvergenceReport object.
Needed to prevent an exception. Apparently both the value and usage
sting must match each time we register a parameter no matter
whether the template parameters differs or not.
- use imported target for linking
- use separate damaris cmake script
- handle HAVE_DAMARIS config variable in the usual way
fixing issues when user does not provide an outputDir via a command_line
avoid adding damaris's command lines when we dont have damaris
This is necessary because when using Damaris we must have the output dir ready
when we initialize Damaris. In the existing code, this was ensured only
from the setupLogging() call.
Damaris initialization is added after InitMpi but before starting the simulation. Damaris will invoke a separate core for writing in
parallel and leave the rest of cores for the simulator. The main changes are in main where start_damaris and then in eclwriterm where
we use damaris to output the PRESSURE. To test Damaris one can use --enable-damaris-output=true and to use parallel HDF5 one can use
--enable-async-damaris-output=true (false is the default choice)
This is only instantiated for two-phase gas/oil and for 3-phase blackoil.
Runtime safeguards have been added to avoid the mistake of running with
a simulator combination that silently ignores DIFFUSE.
This is one of several PRs that will rename some command line parameters for greater consistency, as agreed with Atgeirr in #4096
Here we rename FlowNewtonMinIterations to NewtonMinIterations
Moving it ensures it is also done for restart simulation runs. Without
this, the well model will actually initialize the IQs, but only for
owned cells, not overlap/ghost cells, which caused parallel failure.
this is used transferring ownership of setup structures to the
simulator. drop all the flowEbosXXX set deck methods and use the
generic vanguard. also means various structs that were only passed
in the blackoil simulator are now passed in all simulators.
also shut unsolvable wells during iterations as default
Note however that getWellConvergence return false if a well is unsolvable
i.e. it will cut the timestep and redo the iteration if the well is unsolvable
and only shut it completly if the time step is below the
MinTimeStepBeforeShuttingProblematicWellsInDays (default 0.01 days)
Adding explicit input specification of water-gas ratio (RVW) and RVW output plus simulator gas-water system with salt precipitation and water evaporation for
If no cell has a valid corner-point geometry, typically caused by
using GDFILE to read non-finite data such as all ZCORN = -1.0E+20,
then we must not attempt to generate a grid structure. If we do, we
will typically just fail somewhere deep down in the corner-point
processing code and generate a diagnostic message that's hard to
decipher.
With this commit we instead output a diagnostic message of the form
Failed to create valid EclipseState object.
Exception caught: No active cell in input grid has valid/finite cell geometry
Please check geometry keywords, especially if grid is imported through GDFILE
This may not be a lot better than the original diagnostic
Processing grid
flow: ${ROOT}/opm-grid/opm/grid/cpgpreprocess/preprocess.c:768: is_lefthanded: Assertion `! searching' failed.
Aborted (core dumped)
but does at least suggest that the grid data may be faulty.
There was a fallout when skipping concating these files to the
default ones (PR #1708). We should also have deactivated creating
these files at all. As a result these files appeared in all parallel
runs that were aborted.
This change now prevents creating and logging to these files in
parallel in a default
run (--enable-parallel-logging-fallout-warning=false).
Developers can still activate logging and concating to see whether
everything is only logged on the io process by passing
--enable-parallel-logging-fallout-warning=true.
Closes 3725
This allows for testing the simulator with the artifically split communicator,
in order to verify that there is no inappropriate usage of the world communicator.
In some models there will be only a few cells with rather high pore volume representing the numerical aquifer. Including them (as non-violated cells) in the ratio will make the ratio lower than without the aquifer there and might lead to accepting vectors as converged when they should not.
