The performance would be worse than dyldBut the overhead should be ignorable0.01sec*5 for GCC, 0.17sec*2 for clang