JovanGerb
JovanGerb
!bench I modified the code, expecting it so be a little bit slower (or fail).
Files in the namespace `Mathlib.LinearAlgebra.AffineSpace`, and files containing the name `AddTorsor` have all sped up by about 0.5% due to this change :) The reason is that the `outParam` label...
!bench
!bench
!bench Wow, that last run was really fast 🏎️ Hopefully this one is even faster
!bench
!bench I made a mistake in the previous version, so let's try again
!bench The improvement here is that the canonical instance optimization is applies more often.
!bench I think I fixed the loop-detection to recognize more non-loops