Re: [Ffmpeg-devel] [RFC] benchmark on athlon + intel xeon + Apple G4 + intel p3

Re: [Ffmpeg-devel] [RFC] benchmark on athlon + intel xeon + Apple G4 + intel p3

[Roy Sigurd Karlsbakk]

>Intel Pentium III Cu-mine 850MHz (on busy system)
>1.  bench: utime=87.380s
>2.  bench: utime=87.390s
>3.  bench: utime=87.440s
>avg bench: utime=87.403s

humtitim...
trying out on developer.skolelinux.no - a dual P III 997.5MHz (according to /proc/cpuinfo)
machine:

1. bench: utime=75.790s
2. bench: utime=76.570s
3. bench: utime=76.330s
avg. bench: utime=76.230s

>> K7: Hz*s/1000: 76.963
>> G4: Hz*s/1000: 80.223
>> P4: Hz*s/1000: 93.932
>  P3: Hz*s/1000: 74.293
P3(2): Hz*s/1000: 76.094

strange... this shows marginally lower performance on mine than yours, but it still beats the
lot, including the Athlon. Can anyone explain this???????????

I beleive there was some talk on lkml about some sort of P4 optimisation (interrupt stuff) that
wasn't enabled on 2.4, but comes along in 2.5. Any ideas??? (I cc:ed the lkml on this one)

>Hmm, P3 seems pretty fast in comparison and this was not tested on high-end
>system but standard Linux 2.4.18, P3 850, i440BX, 256MB SDRAM, 10GB 4200UPM
>HDD with lots of other apps running at the same time.

The system I was testing on had a load of around 1.2 or something, but given it's a system
with quite nice I/O and two 1GHz CPUs, I don't think it was important.

roy
--
Roy Sigurd Karlsbakk, Datavaktmester

Computers are like air conditioners.
They stop working when you open Windows.