https://gcc.gnu.org/bugzilla/show_bug.cgi?id=125478
--- Comment #2 from Sujay Kothadiya <skothadiya at whileone dot in> --- Following the suggestion to use -fno-vect-cost-model, I re-ran the bisection. I found this first bad commit: g:638320686c462e7fcdd4e7b3d49d090ac480bd18 With this flag, the issue can be reproduced earlier than previously observed, which is consistent with the earlier observation that the previously identified commit may only have affected vectorization costing.
