On 10/26/22 08:24, Andrew MacLeod wrote:
Figured I would ask what you guys think of making ranger the default
for the VRP1 pass now.
With partial equivalences and the other bits I checked in the past few
weeks I'm not aware of much that the legacy VRP pass gets that ranger
doesn't. The only exception to that which I am aware of is the trick
played with the unreachable edges to set global ranges, but that is
done in the DOM passes now anyway... so it just happens slightly later
in the optimization cycle. There is one test case that needs
adjustment for that which was just checking for a mask in DOM2
(gcc.dg/tree-ssa/pr107009.c). At this point I have not aware of
anything that Id be concerned about, and the testsuite seems to run
cleanly.
We could change the default now and see if any issues show up, giving
us a chance to address them. The code base has been well exercised for
a while so risk should be low. We could also reduce code size by
stripping out unneeded code if we so desired.
Or we could leave things as they are for one more cycle. My
preference would be to make the switch now and let it play out. Thoughts?
I'd rather switch now than wait another cycle. We've got mass rebuilds
coming up in a few months which should give us a good indicator if
there's any notable issues that need to get resolved.
Jeff