Not writing binary output file produces (parasitic?) warning messages

I am using DualSPHysics v4.2.058 (04-06-2018) to run two instances of the default dam-break test case. The configuration file (xml) is the same.

When launching the executable DualSPHysics4.2_linux64 I use the `-stable` option to ensure that repeated simulations give the same results. In the one run I want to have output files (default baseline), while in the other I do not want to read out to output files (flag `-sv:none`). I obtained this information from the help file or `DualSPHysics4.2_linux64 -help`.

It is true that the simulation with the `-sv:none` flag does not produce output files. Unexpectedly though, this simulation throws warnings that do not show up in the baseline simulation.

I observe this from the Run.out file. For example, in the baseline situation at output time steps 68 onwards is

Part_0068 0.680018 38024 523 6174.06 18-02-2019 18:57:21
Particles out: 2 (total: 2)
Part_0069 0.690005 38543 519 6137.49 18-02-2019 18:57:08
Particles out: 18 (total: 20)
Part_0070 0.700013 39063 520 6131.97 18-02-2019 18:56:56
Particles out: 61 (total: 81)
Part_0071 0.710007 39576 513 6056.32 18-02-2019 18:56:44
Particles out: 47 (total: 128)
...
while with the flagged executable, the output is

Part_0068 0.680018 38024 523 6118.74 18-02-2019 17:02:02
Particles out: 2 (total: 2)
*** WARNING: More than 1% of current fluid particles were excluded in one PART (t:0.680018, nstep:38023)

Part_0069 0.690005 38543 519 6077.51 18-02-2019 17:01:50
Particles out: 18 (total: 20)
*** WARNING: More than 2% of current fluid particles were excluded in one PART (t:0.690005, nstep:38542)

Part_0070 0.700013 39063 520 6072.79 18-02-2019 17:01:38
Particles out: 61 (total: 81)
*** WARNING: More than 5% of current fluid particles were excluded in one PART (t:0.700013, nstep:39062)

Part_0071 0.710007 39576 513 5997.37 18-02-2019 17:01:25
Particles out: 47 (total: 128)
*** WARNING: More than 10% of current fluid particles were excluded in one PART (t:0.710007, nstep:39575)
...
The percentage of excluded particles increases continously. In the end report the number of excluded particles is the same for both simulations.

More side information. This issue is reproducible. The same oddity arises when the save flag is `-sv:vtk`. Other flags active in both simulations are `-gpu:0 -svres -posdouble:0`.

Question to the developers: would you please triage this error and provide clues/workaround to sort this out? Perhaps a revised source code?

Although it looks like just a cosmetic glitch, it would be preferrable to clarify whether warning is meaningful at all and, in case it is not, obtain a clean report file.

The guarantee that I need is that the simulations are precisely the same in full compliance with the `-stable` flag.

Thanks in advance for dealing with this.
Sign In or Register to comment.