Using the GPU api for general parallel operations?

I was going to edit my previous post but I accidentally deleted it so made a new one.

I want to make a remeshing addon that uses the GPU api to do the number crunching.

I already confirmed that its possible to encode the mesh data on textures and do the computations on an offscreen buffer using a fragment shader, but I have no idea of how reliable this is.

One problem I’m running into is that my entire computer will freeze for a few seconds if I try to do a ping pong buffer with two offscreens, it works but its still really out of my comfort zone to debug these GPU related issues.

Is there anything important I need to consider for doing this sort of thing with the GPU api?