OPM
Support UDA and restart for RESV control in GCONPROD
The single test failure is to be expected (due to now writing the RESV value in GCONPROD).
The single test failure is to be expected (due to now writing the RESV value in GCONPROD).
Thanks. The simulation results are untouched by the PR, while it is indeed more restart data was written in the RST file,
I am not following where the value 1.0000000e+20 comes from, @vkip , do you have a understanding?
SGRP: Restart, sequence 11
Fails for 2 entries
Largest absolute error: 1.0000000e+20
Largest relative error: 1.0000000e+00
Others probably can review the PR more efficiently than me, so I am delegating this to @bska . I will review this if it does not work.
am not following where the value
1.0000000e+20comes from
1.0e+20 is a sentinel value that, in SGRP and a few other arrays, means "no limit". Prior to this PR we did not emit targets/limits for GCONPROD/RESV so we would always write this sentinel value for the RESV limit at the group level. This PR ensures that we write the correct value of 1000 for the RESV limit. In other words, the regression failure is expected.
Prior to this PR we did not emit targets/limits for GCONPROD/RESV so we would always write this sentinel value for the RESV limit at the group level. This PR ensures that we write the correct value of 1000 for the RESV limit. In other words, the regression failure is expected.
Thanks for pointing to the following,
'F' 'RESV' 4* 'RATE' NO 5* 1000 /
I did not spot it. But I did not see the 1000 showing up in the regression failure report.
https://ci.opm-project.org/job/opm-common-PR-builder/8384/testReport/(root)/mpi/compareECLFiles_flow_SPE9_CP_GROUP_RESV/
or it is too small against 1.e+20 to show up?
I did not see the 1000 showing up in the regression failure report.
It's there, but you need to read the full comparison, not just the summary at the end:
[...] Unified restart files, sequence 11 Comparing SEQNUM ... done. Comparing INTEHEAD ... done. Comparing LOGIHEAD ... done. Comparing DOUBHEAD ... done. Comparing IGRP ... done. Comparing SGRP ... Keyword: SGRP, origin Restart, sequence 11 Value index = 147 (first value, second value) = (1e+20, 1000) Keyword: SGRP requires strict tolerances. [...]
I did not see the 1000 showing up in the regression failure report.
It's there, but you need to read the full comparison, not just the summary at the end:
[...] Unified restart files, sequence 11 Comparing SEQNUM ... done. Comparing INTEHEAD ... done. Comparing LOGIHEAD ... done. Comparing DOUBHEAD ... done. Comparing IGRP ... done. Comparing SGRP ... Keyword: SGRP, origin Restart, sequence 11 Value index = 147 (first value, second value) = (1e+20, 1000) Keyword: SGRP requires strict tolerances. [...]
That is correct. Thanks.