https://gcc.gnu.org/bugzilla/show_bug.cgi?id=120231
--- Comment #11 from Jakub Jelinek <jakub at gcc dot gnu.org> --- Created attachment 61583 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=61583&action=edit gcc16-pr120231.patch Untested patch for the float -> int and int -> float casts and their reverse ops. Unfortunately it regresses FAIL: g++.dg/tree-ssa/loop-split-1.C scan-tree-dump-times lsplit "loop split" 1 FAIL: gfortran.dg/inline_matmul_16.f90 -O scan-tree-dump-times optimized "_gfortran_matmul" 1 tests, so I'll need to study those up tomorrow.
