Any micro performance gains from a function call will become rounding errors compared to the extra cost of any arithmetic operation on a dynamically-typed langu

Any micro performance gains from a function call will become rounding errors compared to the extra cost of any arithmetic operation on a dynamically-typed language. It may be an interesting article for describing how JIT and some late-optimization techniques can produce optimized code given the right optimization preconditions, but I fail to see how the title is supposed to describe the contents of the article. Last, but not least, performance-sensitive C code tends to be statically linked.