Inner-loop operations

The preceding basic optimizations complete our work with the atomic software. To further explore the issue of floating point performance in general, this problem allows you to evaluate for yourself, on your own computer, the compute rate and relative cost of different operations in the innermost loop.

To help you with the problem, Appendix B contains a function timer() which computes the FLOP rate for the various subroutines which you will write; a main() program which prints out various information; and a function mult() as an example of a routine to be timed. At present, much of the main() program is commented out. By the end of this problem set, you should have all of main() program functioning.
Note: Because of operating system dependencies in timing routines, those using the Windows compiler should replace timer() from Appendix B with the routine in Appendix C.