This adds some extra costs for reverse shuffles under AArch64, filling in the i16/f16/i8 gaps in the cost model. Differential Revision: https://reviews.llvm.org/D124786
nocallback
-analyze -divergence
-passes='print<divergence>'