https://github.com/paulwalker-arm commented:
I could not spot any tests for the `<2 x float>` vector functions? Hopefully it's as easy as adding extra RUN lines to veclib-function-calls.ll and veclib-intrinsic-calls.ll. I'm thinking variants of LIBMVEC-NEON that use `-force-vector-width=2`. If this works then it's worth changing the existing LIBMVEC-NEON RUN lines to use `-force-vector-width=4` so that both variants are locked in. https://github.com/llvm/llvm-project/pull/143696 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
