It's an interesting approach. A hybrid software with limited hardware change is interesting.
> Fil-C will get faster even without architecture help
Only to a point. You'll plateau eventually since you have checks on every dereference.
> Fil-C will get faster even without architecture help
Only to a point. You'll plateau eventually since you have checks on every dereference.
That’s what I’m working on now. Should have some early results soon.
That plateau might not be any different than the CHERI plateau, since at the microarch level, CHERI will have more checks since it cannot benefit from a compiler’s static reasoning about redundant check elimination.