[Clang] Correctly enable the f16 type for offloading

Summary:
There's an extra argument that's required to *actually* enable f16
usage. For whatever reason there's a difference between fp16 and f16,
where fp16 is some weird version that converts between the two. Long
story short, without this the math builtins are blatantly broken.

Author: jhuber6

clang/lib/Basic/Targets/NVPTX.h

@@ -75,6 +75,8 @@ class LLVM_LIBRARY_VISIBILITY NVPTXTargetInfo : public TargetInfo {
 
   ArrayRef<Builtin::Info> getTargetBuiltins() const override;
 
+  bool useFP16ConversionIntrinsics() const override { return false; }
+
   bool
   initFeatureMap(llvm::StringMap<bool> &Features, DiagnosticsEngine &Diags,
                  StringRef CPU,