Wrong Results of examples/freeSurface (v1.8.1)

[ZiwenZhang]

I recently ran the five examples in examples/freeSurface from OpenLB v1.8.1 using the onlyGPU parallelization mode on an RTX 3060. I found that only FallingDrop2D and FallingDrop3D work properly, while in the other three cases, both mass and epsilon remain completely unchanged.

This confuses me, since I did not modify any of the default settings. I’m wondering if this might be a bug introduced in a recent update to a specific module. However, the fact that the FallingDrop cases run correctly makes me unsure.

Could you please let me know whether the simulations of the other three cases run normally on your side, or where the problem might lie?

Thank you very much.

[Maximilian]

Hi ZiwenZhang, I’ve looked into it and I can so far confirm that there is some error when executing the freeSurface examples on GPU. Running them on CPU gave me correct results.

You mentioned that the fallingDrop examples run correctly on your end. Can you tell me how the visualization of fallingDrop2d looked for you?

For me, the fallingDrop examples show movement of the liquid, but the behaviour is faulty: The liquid pool on the bottom ‘curls up’, when it is supposed to ‘splash’ to the sides.

Other examples such as rayleighInstability or breakingDam show no movement at all.

Maximilian

[ZiwenZhang]

Hi Maximilian, Thank you for your reply. The situation is as same as yours: The liquid pool on the bottom ‘curls up’, and other examples such as rayleighInstability or breakingDam show no movement at all. It seems there might be some bugs.

[Author]

Posts