Pierre added a comment. In D60763#1529279 <https://reviews.llvm.org/D60763#1529279>, @Nicola wrote:
> A bit late to the review, but I've noticed a couple of issues with some of > the implemented builtins: > > - The fmin/fmax builtins are defined twice for scalar types, does this create > problems in overload resolution when using them? > - The convert_ builtins don't have support for half types (which is present > in the opencl-c.h header. Is that intended? Both your comments are rights. Some builtin functions are currently a faulty/missing. I am currently updating the functions so that they match the OpenCL C specification. convert_half will be present, and the way builtin functions are defined in the .td file will change a bit. The new version should be available by the end of the week. Repository: rL LLVM CHANGES SINCE LAST ACTION https://reviews.llvm.org/D60763/new/ https://reviews.llvm.org/D60763 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
