Seems a bug in test_2d_eulerian_supersonic_flow_new_BC

[Xiangyu-Hu]

Test project C:/build
Start 25: test_2d_eulerian_supersonic_flow_new_BC
1/1 Test #25: test_2d_eulerian_supersonic_flow_new_BC ...***Failed 247.04 sec
Particle relaxation was set to default (0).
Particle reload from input file was set to 1.
Generate regression test data was set to default (0).
State recording was set to 0.
Restart inactivated, i.e. restart_step (0).
N=0 Time = 0.003095238 dt = 0.003095238
N=500 Time = 1.426021891 dt = 0.002668712
N=1000 Time = 2.787473310 dt = 0.002738317
N=1500 Time = 4.179992051 dt = 0.002821461
N=2000 Time = 5.594371866 dt = 0.002884705
N=2500 Time = 7.021402524 dt = 0.002904886
N=3000 Time = 8.481243027 dt = 0.002912009
N=3500 Time = 9.949600680 dt = 0.002946277
N=4000 Time = 11.305570902 dt = 0.002914381
N=4500 Time = 12.701695884 dt = 0.002951483
N=5000 Time = 14.182943269 dt = 0.002965377
N=5500 Time = 14.757984083 dt = 0.000583324
N=6000 Time = 14.977852726 dt = 0.001506186
N=6500 Time = 16.033056754 dt = 0.002840551
N=7000 Time = 17.405570922 dt = 0.002758674
N=7500 Time = 18.657846138 dt = 0.002815529
N=8000 Time = 20.063502283 dt = 0.002797121
N=8500 Time = 21.471543551 dt = 0.002632537
N=9000 Time = 22.892970999 dt = 0.002766666
N=9500 Time = 24.281016416 dt = 0.002743484
N=10000 Time = 25.709847419 dt = 0.002883958
N=10500 Time = 27.143265868 dt = 0.002975662
N=11000 Time = 28.119770189 dt = 0.001727697
N=11500 Time = 28.795259872 dt = 0.000989920
N=12000 Time = 29.465842161 dt = 0.001940631
N=12500 Time = 30.315705353 dt = 0.001829291
N=13000 Time = 31.366627826 dt = 0.002304918
N=13500 Time = 32.033029155 dt = 0.001323722
N=14000 Time = 32.748799287 dt = 0.001551817
N=14500 Time = 33.276035170 dt = 0.002634898
N=15000 Time = 34.113414821 dt = 0.002040318
N=15500 Time = 35.173542284 dt = 0.001210626
N=16000 Time = 36.079390856 dt = 0.001161615
N=16500 Time = 36.950305276 dt = 0.001572580
N=17000 Time = 37.594424506 dt = 0.001451106
N=17500 Time = 38.345759972 dt = 0.001394039
N=18000 Time = 39.223442823 dt = 0.001516370
Total wall time for computation: 246.949553100 seconds.
MaximumSpeed[0] in Snapshot_500 is beyond the exception, and difference is 153.253523574
MaximumSpeed[0] in Snapshot_1000 is beyond the exception, and difference is 1931952551.898786783
MaximumSpeed[0] in Snapshot_1500 is beyond the exception, and difference is 1938725614.755589962
MaximumSpeed[0] in Snapshot_2000 is beyond the exception, and difference is 1945494983.009193897
MaximumSpeed[0] in Snapshot_2500 is beyond the exception, and difference is 1945984443.356199026
MaximumSpeed[0] in Snapshot_3000 is beyond the exception, and difference is 1943943415.744412661
MaximumSpeed[0] in Snapshot_3500 is beyond the exception, and difference is 1941319130.624794006
MaximumSpeed[0] in Snapshot_4000 is beyond the exception, and difference is 1939071393.784265280
MaximumSpeed[0] in Snapshot_4500 is beyond the exception, and difference is 1947290793.873203516
MaximumSpeed[0] in Snapshot_5000 is beyond the exception, and difference is 1947922931.702103853
MaximumSpeed[0] in Snapshot_5500 is beyond the exception, and difference is inf
MaximumSpeed[0] in Snapshot_6000 is beyond the exception, and difference is inf
MaximumSpeed[0] in Snapshot_6500 is beyond the exception, and difference is inf
MaximumSpeed[0] in Snapshot_7000 is beyond the exception, and difference is inf
MaximumSpeed[0] in Snapshot_7500 is beyond the exception, and difference is inf

It seems that some time the code go wrong.

[Xiangyu-Hu]

Is the bug found?

[ZhentongWang]

Is the bug found?

I checked the numerical results, and the results are fine.
The problem seems happening in the regression file. I have created the pull request where I use a different regression strategy to create the regression file.

[ZhentongWang]

Is the bug found?

some time, its other's case failed in the Linux-test, and I will re-run this again.

[Xiangyu-Hu]

Is the bug found?

some time, its other's case failed in the Linux-test, and I will re-run this again.

Yes. But the optimization case is not certain by itself. But the Eulerian case should be certain.

[ZhentongWang]

Is the bug found?

some time, its other's case failed in the Linux-test, and I will re-run this again.

Yes. But the optimization case is not certain by itself. But the Eulerian case should be certain.

Currently, the Eulerian supersonic flow case can pass.

[Xiangyu-Hu]

So no big is found?

[Xiangyu-Hu]

How can we explain the issue list above?

[ZhentongWang]

So no big is found?

I have tried many times to simulate this case, and the results (including velocity contour) seem correct and convincing.
Therefore, I don't think there is a bug in the simulation.

[ZhentongWang]

How can we explain the issue list above?

In Linux: I have tried many time and the output is stable as follows:

In Window: also tried many times and the output is as follows:

I have noticed that the output that you uploaded shows there was many dt very small, but I haven't met this phenomenon during the many simulations I have done.

[Xiangyu-Hu]

@ZhentongWang may be the problem is related to particle generation. For you many simulation did you re generate particle every time?

[ZhentongWang]

@ZhentongWang may be the problem is related to particle generation. For you many simulation did you re generate particle every time?

Yes, I use particle relaxation every time.
Here, some results in Linux show as follows:

Here, some results in Windows show as follows:

The dt in these results also seems stable.