execute has to fork, call execv in the child, wait for the child, and save output and status code in memory. Using execv directly makes your compiler become the program you're running, without ever creating an additional process. Not only is this more efficient, execute will simply not work for programs with large (or infinite) output.
