Tagging subscribers to this area: @JulieLeeMSFT, @jakobbotsch
See info in area-owners.md if you want to be subscribed.

As per discussion in Discord, if you want to land something like this you will need to make OptRepeatCount robust alone without any changes. Currently, DOTNET_OptRepeatCount=x will trigger lots of asserts everywhere, you might avoid them if you limit your opts to some specific loop-unrolling cases but it will only mean that it might still crash somewhere else outside of BCL.

We also expect nice diffs from OptRepeatCount alone (as well as massive throughput regressions due to SSA/VN re-computations).

as well as massive throughput regressions

Well that doesn't sound ideal ;-)

as well as massive throughput regressions

Well that doesn't sound ideal ;-)

Yep, but maybe it might be allowed for some very specific cases or NativeAOT where we can afford slightly longer compilation 🙂

Closing as not beneficial.