> I think this gets at Wirth's law a bit. I think Proebsting's law is also relevant. Compilers did an impressive job of optimization in the 90s, and while yes

> I think this gets at Wirth's law a bit. I think Proebsting's law is also relevant. Compilers did an impressive job of optimization in the 90s, and while yes there have been advancements in the last 18 years, it's foolish to over-estimate the progress. Also, considering how much hardware changes have improved performance (speculative execution, register renaming, etc...), it wouldn't surprise me if many compiler optimizations have become less important with time. For instance, peep-hole instruction scheduling has to matter less than it once did.