Once #111659 gets landed, we need to investigate how the support matrix for clang and gcc looks like for the complex variant of float16 and float128. Once we have that information, add header guards for them accordingly and add support for them in FPMatcher.