Windows Linux differences in msl32

[Christian Schubert]

Hi,

we currently have three differences between Windows and Linux:

• Modelica.Fluid.Examples.InverseParameterization Linux complains two times fewer about an assert while solving a nonlinear system -> I would ignore that
• Modelica.Magnetic.FundamentalWave.Examples.BasicMachines.AIMS_Start It gives "Files Equal" under Windows but "Files not equal" under Linux
• Modelica.Mechanics.MultiBody.Examples.Constraints.RevoluteConstraint fails to translate under Linux but gives "Files Equal" under Windows.

I would disable the last two tests for now. But someone should look into it as soon as possible.

[Adrian Pop]

I did fixed all these tests. At least TEST_CLANG seems to be 100% again.
We'll see about TEST later on.

I had to change the matching algorithm for RevoluteConstraint to make it work in Linux.
I reported RevoluteConstraint issue in #2405.

Cheers,
Adrian Pop/

[Martin Sjölund]

Modelica.Fluid.Examples.InverseParameterization Linux complains two times fewer about an assert while solving a nonlinear system -> I would ignore that

I would make it so we don't spam that message so much (in ThermoPower there is one model that sends this messages so much the log is 57MB)

Modelica.Magnetic.FundamentalWave.Examples.BasicMachines.AIMS_Start It gives "Files Equal" under Windows but "Files not equal" under Linux

What is the error tolerance? Sometimes the error is quite small.

[Adrian Pop]

For Modelica.Magnetic.FundamentalWave.Examples.BasicMachines.AIMS_Start.mos the file in subversion that you commited had "Files not equal". I updated it with running the model in Linux and that gave "Files Equal!". So I think you have these reversed somehow (maybe it fails in Windows and works in Linux?).

[Christian Schubert]

Yeah, you're right, I mixed it up.
I would also agree with Martin. The warning on

|                 | |       | | While solving non-linear system an assert was called.
|                 | |       | | The non-linear solver tries to solve the problem that could take some time.
|                 | |       | | It could help to provide better start-values for the iteration variables.
|                 | |       | | For more information simulate with -lv LOG_NLS

is not helpful for an end-user, sine I would not know what to do about it.
It would only be interesting to know if the simulation terminates due to nonlinear solver failure, that asserts have occured.
Nevertheless, this output may be printed if the -lv=... option is used to debug the simulation.

[Lennart Ochel]

Modelica.Mechanics.MultiBody.Examples.Constraints.RevoluteConstraint is working fine.

[Lennart Ochel]

Modelica.Magnetic.FundamentalWave.Examples.BasicMachines.AIMS_Start produces "Files Equal" for Windows and Linux.

[Lennart Ochel]

Modelica.Fluid.Examples.InverseParameterization produces "Files Equal" on Windows and Linux, but it still generates different warnings for both operating systems.

[Martin Sjölund]

This ticket was not closed for 1.9.1, which has now been released. It was batch modified for milestone 1.9.2 (but maybe an empty milestone was more appropriate; feel free to change it).

[Martin Sjölund]

Milestone changed to 1.9.3 since 1.9.2 was released.

[Martin Sjölund]

Moved to new milestone 1.9.4

[Adrian Pop]

The test results are the same but the warnings about NLS convergence are not.

[Martin Sjölund]

Milestone renamed

[Martin Sjölund]

Milestone renamed