[MPlayer-dev-eng] New audio filter layer finished

[Michael Niedermayer]

Hi

On Thursday 03 October 2002 10:55, Anders Johansson wrote:
> > > I examined the assembly output and concluded that the code was
> > > better optimized when I used 16bit aritmetics... It is a bit wierd now
> > > when I think about it, It should be the same. I have to look at it
> > > some day...
> >
> > did u benchmark it? 16-bit tends to be slower than 32-bit on 32-bit CPUs
>
> No guess I should have - was quite new to linux and x86 programming
> when I wrote the code. btw what SW do you use for doing that type of
> bench marking?
static inline long long rdtsc()
{
	long long l;
	asm volatile(	"rdtsc\n\t"
		: "=A" (l)
	);
	return l;
}

this will return the exact number of cpu cycles which have passed since some 
point (system boot or something like that)
btw, this works only on x86 (and pentium or newer IIRC)

[...]

Michael