I have to believe the same is true under Visual Studio 2019, but
offhand don't know how to prove that. I understand Steve uses PGO to
build the python.org Windows release, but I've never done that - the
"Release build" configuration I get from Github does not use PGO, and
the code I get from my own "release builds" is equally slow for all
divisors (and the generated assembler is obviously not trying to
special-case anything).