Very impressive! Not long ago this would have taken a very strong GPU programmer with extensive physics simulation knowledge at least a year!
And most likely there are still good perf gains on the table: getting the maximum out of GPU sim usually requires a non-trivial
Fable and I rewrote Box3d entirely in native cuda. It's about 30x faster on my GPU compared to my CPU (depending on the situation).
This was my hardest benchmark. I still had to be there, but soon I won't. Cost me about 200 dollars
