[libc][math][c23] Improve performance of fadd using double arithmetic

libc/test/src/math/performance_testing/CMakeLists.txt

@@ -475,6 +475,18 @@ add_perf_binary(
     libc.src.math.fminimum_numf16
   COMPILE_OPTIONS
     -fno-builtin
+    )
+
+add_perf_binary(
+  fadd_perf
+  SRCS
+    fadd_perf.cpp
+  DEPENDS
+    .binary_op_single_output_diff
+    libc.src.math.fadd
+    libc.src.__support.FPUtil.fp_bits
+  COMPILE_OPTIONS
+    -fno-builtin
 )
 
 add_perf_binary(

libc/test/src/math/performance_testing/fadd_perf.cpp

@@ -0,0 +1,23 @@
+//===-- Performance test for the fadd function ----------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include "BinaryOpSingleOutputPerf.h"
+#include "src/math/fadd.h"
+
+static constexpr size_t DOUBLE_ROUNDS = 40;
+
+float fadd_placeholder_binary(double x, double y) {
+  return static_cast<float>(x + y);
+}
+
+int main() {
+  BINARY_OP_SINGLE_OUTPUT_PERF_EX(float, double, LIBC_NAMESPACE::fadd,
+                                  fadd_placeholder_binary, DOUBLE_ROUNDS,
+                                  "fadd_perf.log")
+  return 0;
+}