Always best to go direct to the source on these things (emphasis mine):

NOTE: The meaning of this PR, potential perf impact and the problem it tries to solve is being grossly misrepresented on other sites.

The goal of this refactor is to optimize for cases where games (Starfield in particular) uses advanced ExecuteIndirect in very inefficient ways.

[…]

To be clear, the gains expected here are very minute. Single percent range to pop some final bubbles that Mesa didn’t clean up on its own. The real gains come from recent Mesa patches on main.