Use "restrict" to achieve even higher performance in benchmarks

[dumblob]

I see you're doing plenty of interesting benchmarks. I was thinking that according to my understanding, Morty is pretty much "aliasing-safe" and thus it might greatly benefit from using restrict (either in the virtual machine itself or in the transpiled Morty to C).

Have you thought about using restrict?

[mobarski]

I haven't thought about it, but it seems like a good direction to investigate.
My current benchmarking showed me that current VM optimizations can improve performance by around 100%, which is nice, but it's not as significant as I thought it will be.
Transpilation on the other hand gives 90% of the C performance (when doing stack tracking in the Morty compiler, and using gcc -O3 ) or 50% when implemented naively. So my current direction is to treat the VM as a permacomputing solution and transcompilation as a performant / energy-friendly solution.
I'm trying to be able to transpile Morty VM cell-code (not only Morty source code) into target languages.