https://gcc.gnu.org/bugzilla/show_bug.cgi?id=89271
--- Comment #14 from Segher Boessenkool <segher at gcc dot gnu.org> --- Cool results so far! +FAIL: gcc.target/powerpc/p9-dimode1.c scan-assembler-not \\mmtvsrd\\M p9_plus_1: .LFB1: .cfi_startproc - vspltisw 0,1 - vupklsw 0,0 + li 9,1 + mtvsrd 0,9 The new code is actually cheaper on p9. The test is wrong?