Gases don't diffuse properly

Interesting bug.

Some updates from further experimentation:

• The diffusion does go at a reasonable speed if there's a sufficient pressure difference.  For example, the boundary between vacuum and oxygen moves at a reasonable rate as they combine, but the slowness occurs as the low-pressure oxygen only gains a few mg per volume per timestep.
• On equalizing pressure: it looks like the game is trying to equalize the pressure between gases, but instead it's equalizing density.  Presumably it will eventually be fixed by giving each gas a correct molecular mass and equalizing (density * temperature / molecular mass)
• On that note, perhaps chlorine is floating above CO2 because it's programmed as Cl instead of Cl2.  In the gases' information boxes, oxygen and carbon dioxide are correctly labeled as O2 and CO2, but hydrogen and chlorine are incorrectly labeled H and Cl instead of H2 and Cl2.

I decided to test the right bias of CO2 when mixed with oxygen after the release notes for one of the development builds stated "Gas displacement should no longer have a directional bias".  I honestly have no clue what was actually fixed, but the tendency of heavier gasses to settle to the right was not it.  I made two identically sized thermally insulated long rooms and placed CO2 and Oxygen on each side using the debug tools with vacuum in between (this was done while paused, the image shows a few seconds afterwards, where some mixing had occurred rapidly.)

After two cycles the CO2 in the bottom room had switched place to the right side, hugging the wall.

2 hours ago, Hornwitser said:

I decided to test the right bias of CO2 when mixed with oxygen after the release notes for one of the development builds stated "Gas displacement should no longer have a directional bias".  I honestly have no clue what was actually fixed, but the tendency of heavier gasses to settle to the right was not it.  I made two identically sized thermally insulated long rooms and placed CO2 and Oxygen on each side using the debug tools with vacuum in between (this was done while paused, the image shows a few seconds afterwards, where some mixing had occurred rapidly.)

After two cycles the CO2 in the bottom room had switched place to the right side, hugging the wall.

I think it is much more faster to calculate the gases this way, you don't want to use your processor just calculating the gases. Maybe they will solve this later with a bunch of other optimalization (pre-release).