https://gcc.gnu.org/bugzilla/show_bug.cgi?id=127203
--- Comment #3 from Andrew Macleod <amacleod at redhat dot com> --- The issue is that its tripping up the fix for PR 126531. That fix added soe checking for abnormal edges during the range_from_dom saerch, and this testcase in particular it causing the number of dominated_by_p calls to rise from something like 30 million calls to 230 million calls... Its doing a significant larger amount of work now as adding these SSA_NAMES that are used in an abnormal PHI somewhere triggers a lot more detection of such edges. I have a potential fix or two in mind... The current code can be made more efficient, but Im considering a different change in approach. just wanted to update the status of the PR.
