[libc-commits] [libc] [libc][math] Optimize misc basic math operations with builtins when available (PR #99037)

[via libc-commits]

================
@@ -0,0 +1,14 @@
+_Float16 try_builtin_fmaxf16(_Float16 x, _Float16 y) {
+  return __builtin_fmaxf16(x, y);
----------------
lntue wrote:

shall `f16` variants be tested separately?

https://github.com/llvm/llvm-project/pull/99037