https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83726
--- Comment #8 from Andrew Pinski <pinskia at gcc dot gnu.org> --- (In reply to Steve Ellcey from comment #7 > pr83081.f90 looks like a new test so it may not actually be a regression > related to this patch, but the first two are. pr79041-2 generates: > > t: > mov x0, 0 > mov x1, 65536 > ret > > with the patch rather than: > > t: > adr x0, .LC0 > ldp x0, x1, [x0] > ret > .size t, .-t > .align 4 > .LC0: > .xword 0 > .xword 65536 The first code generation is better than doing a load. That is doing two mov is better than doing an adr/ldp in terms of size and even ldp could miss the cache while the two mov are independent and can most likely be executed on the same pipeline.
