When measuring execution time with
real time like clock() also interrupting
background processes are measured,
and the result is always too large
and variable.
How can the clean execution time without
interrupting background processes be
measured on a Pentium 4?

Maarten Kronenburg wrote:
> When measuring execution time with
> real time like clock() also interrupting
> background processes are measured,
> and the result is always too large
> and variable.
> How can the clean execution time without
> interrupting background processes be
> measured on a Pentium 4?

Take a look at the RDTSC instruction. However for long instruction
sequences, you're out of luck unless you can convince the OS to let you
turn off the interrupts.
Cheers,
Randy Hyde