Re: time tutorial

[Lutz Sammer <johns98@gmx.net>]

Erik wrote:
> 
>         -Erik <erik@smluc.org> [http://math.smsu.edu/~br0ke]
> 
> The opinions expressed by me are not necessarily opinions. In all
> probability, they are random rambling, and to be ignored. Failure to ignore
> may result in severe boredom or confusion. Shake well before opening. Keep
> Refrigerated.
> 

Your wish is an order,

First to avoid zero times or infinite results, scale your results and
add
a call counter to the result.

	static clock_t last_clock;
	static int clock_counter;
	int new_clock;
	
	new_clock=clock();
	if ( new_clock == last_clock ) {
		return last_clock*100 + clock_counter ++;
	}
	clock_counter=0;
	last_lock=new_clock;
	return last_clock*100;


Second you forget the "tsc" (time stamp counter) register, which is
available atleast
on any ">=586" and alphas.
This gives you the exact time, no overhead (task-switches, calls, ...),
here is the code:

/**
**      Read the tsc (time-stamp counter) register of the processor.
**
**      @return the rtc register
*/
static __inline u_int64_t rdtsc(void)
{
#ifdef __i386__
    u_int64_t rv;

    __asm __volatile(".byte 0x0f, 0x31" : "=A" (rv));
#endif
#ifdef __alpha__
    int rv;

    __asm __volatile("rpcc %0" : "=r" (rv));
#endif

    return rv;
}

Johns
-------------------------------------------------------------------------
ALE Clone: Free W*rCr*ft II source: http://yi.org/johns