On Fri, 08 Oct 1999 17:12:06 -0700, Monty wrote:
>> > Keep in mind that due to cache alignment (and other) issues, runs of the same
>> > executable can vary in speed by almost half depending on where in the buffer
>> > cache they get placed.
>>
>> I've never found a discrepance of 1/2 in any significant sized program, but
>> I sometimes see variations of +/-5%. (in gprof output etc)
>
>Oh, I knew I should have mentioned that this was Linux :-) I'm sure it's
>possible elsewhere, but what I describe I saw under 2.0 Linux kernels. I
>wonder is 2.2 does the same thing....
>
>Monty
heh I also should have mentioned the speeds I stated were of LAME running
in OS/2, and were with nothing else notable running to take any CPU away
from it. And the speeds are generally consistant with what I've seen in
usual day-to-day usage here. YMMV :)
Paul
--
MP3 ENCODER mailing list ( http://geek.rcc.se/mp3encoder/ )
