That was from the developers mailing list. I think the only optimization they're doing is for FOR statements, where you know the bounds on a loop variable at loop start. This is a big win for loops that iterate over an array and do something trivial, like setting the array elements to 0. Subscript check performance is mostly an inner-loop problem.
Oh, I'm curious why you pointed out Go specifically, since that sounds like fairly standard bounds check eliminations that pretty much all optimising compilers can do (compilers built on both LLVM and GCC's infrastructure), and most compilers aren't tied to specific looping constructs for it: it can work with while loops, or for loops, or whatever.
And, my impression is that bounds checks are only avoided in Go with range-based for loops (which are very similar to C++ or Rust iterators, which avoid bounds checks in probably exactly the same way), not a loop over integers with indexing.
