My stress simulation code iteratively solved F=kx. Basically just springs between all adjacent particles (force is proportional to displacement). However, real stress is more complicated than that. Depending on how accurate you want to make it, there's stuff like bending moments, Poisson's ratio, and thermal expansion to consider. My code was adequate for a column (including a hanging column) with vertical load only and no horizontal forces applied, but gave incorrect results for beams and cantilevers. If we're going to have stress simulation in Powder Toy, I'd like it to bear some resemblance to reality.

Here are some pictures and a video of my attempt at stress simulation: https://jacksonmj.minus.com/mOvf4yuI1

I may (or may not) have another go at some point. Don't hold your breath, doing decent stress simulation is about as hard as doing proper moving solids.

@rodo (View Post)
this is late (because of a necro), but i'll still respond.

tpt++ is a little slower, I don't know why. On a completely empty screen after just starting, I get 380 fps in my mod (tpt) and 250 fps in tpt++. That's a 130 fps difference, but it's not noticeable at all. When there are particles, it's closer to the same fps I think.

The next version will have a button to get your stamps back.

the pressure leak glitch has always existed forever, and hasn't been fixed, or even tried to be fixed. It's really not that important though, you were the first one to ever notice it. The bug even has some uses, Catelite made a cool save to show it off once, and so I don't see a point in fixing it if it can be useful.