Provenance (and to some degree the entire memory model) is the big exception to this. It's a very core part of the IR semantics, and an optimizing compiler without provenance would have to use entirely different methodology to justify many optimizations. I don't believe that anyone has yet been able to demonstrate the feasibility of highly optimizing compiler without provenance.
For example, LLVM IR is used to compile C both with and without strict aliasing. The choice is lowered into metadata applied to individual pointer operations. Similarly with signed integer overflow.
Nim could even have gotten some of this benefit when compiling to C if it had been willing to target a specific set of C compilers, rather than standard C.