Re: Interesting pentium-memcpy results

Ingo Molnar (mingo@pc7537.hil.siemens.at)
Tue, 29 Jul 1997 18:31:56 +0200 (MET DST)


On Tue, 29 Jul 1997, Albert D. Cahalan wrote:

> > +__memcpy_g (void *_to, const void *_from, __kernel_size_t _bytes)
> > +{
> > + if (bytes >= 1024) {
>
> That is what I meant: the test itself adds overhead.

[sorry i misunderstood your sentence.]

i dont think this test adds noticeable overhead. Small constant memcpy's
are already filtered out by the 'builtin_constant' stuff, and thats the
majority. I cant think of anything in the 'few bytes range' that has
variable size memcpy ... maybe i'm missing something.

-- mingo