Messages in this thread | | | Date | Fri, 19 Apr 1996 09:14:11 +0300 (EET DST) | From | Linus Torvalds <> | Subject | Re: Ideas for reducing memory copying and zeroing times (fwd) |
| |
On Fri, 19 Apr 1996, Michael Riepe wrote: > > I guess the fastest (and shortest) way to do that on an 80x86 is: > > xorl %eax,%eax > leal 4096(page_start),%esp > <repeat 1024 times> > pushl %eax > <end repeat>
No, the fastest, and shortest way to do it is
.. dest in %edi .. movl $1024,%ecx xorl %eax,%eax rep; stosl
Which is in fact exactly how linux does it..
Of course, if you have an old x86 chip, that's your problem and you may not get optimal performance, but who expected anything else from old hardware?
(Hint: the above _really_ flies on a PPro. Intel optimized it to do cache-line accesses, it seems. They did the right thing)
Linus
| |