nemanjai added inline comments. ================ Comment at: test/CodeGen/builtins-ppc-vsx.c:131 @@ -118,2 +130,3 @@ +// CHECK: fdiv <2 x double> /* vec_max */ ---------------- wschmidt wrote: > Have you verified that we still generate the xvdiv*p instructions when this > is done end-to-end? I want to be sure this doesn't accidentally scalarize > the code. Yes, I have already confirmed this with the patch applied. Thank you.
Repository: rL LLVM http://reviews.llvm.org/D11184 _______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
