times2
14th May 2002 14:18 UTC
amd vs. intel for avs
hi all
as i'm upgrading my pc i wonder if anyone
out there has used avs on a p4 (i haven't)
and especially if anyone has compared
athlon performance to p4 performance.
i'd especially be interested if anyone
has the ability to compare performance
using my pixelplayer ape (see my 'ape goody' thread).
thanks
cmountford
19th May 2002 00:45 UTC
I'm sorry that I don't know, but I felt it necessary to get your post back near the top. You might PM or e-mail UnConeD your question, though. :(
Just thought it had gone far too long without any replies to your thread, so I posted a reply to make it appear at the top of the thread list. Hopefully some of the people (AVS VJs: I'm talking to you!) who have used AVS on different computer setups will see this post.
Sorry - but my best computer is a P3 450 MHz, so I wouldn't be of help there :(
Adasha
19th May 2002 14:37 UTC
my general experience is that Athlons are better at floating point maths than pentiums so any non-integer calculations should be faster at an equivalent MHz speed.
Having said that I'm sure someone from the pentium camp would be able to supply reasons why they are better, i guess it's down to what you can afford and your personal preference.
times2
20th May 2002 10:45 UTC
athlon vs. p4
i've done a couple of tests....
machine 1: athlon 1400 (non-XP), ali magik, pc2100 ddr
machine 2: p4 xeon dual 1800, i850 dual rambus pc800
machine 3: p3 dual 667, pc800 rambus
preset 1: 2 'pixelplayers', all the slowest settings (oversample, fader on ~50%). playing at 320x240. (
machine 1, 49.9fps.
machine 2, 57 fps.
machine 3, 32 fps.
preset 2: el-vis 'real 3d sentinel' (320x240)
machine 1, 25 fps.
machine 2, 32 fps.
machine 3, 21.3 fps.
this is a pretty damning result for the p4, that machine cost £4000 last week whereas the athlon cost £1400 a year ago!!
BUT when running preset 2, the p4 is running at 25% cpu use (=50% of 1 cpu) whereas the athlon is running at 75%. SO i would say a lot of the difference is down to AVS not being optimised for the p4, and therefore causing pipeline stalls etc.
can anyone else try these tests on their machines?
nb the display settings I used were MAXIMUM windowed performance, DONT wait for retrace, NO pixel doubling. And i used a 32-bit display mode for both (which is far faster for avs). There were no other programs running other than that winamp was playing an mp3.
Atmo
20th May 2002 19:24 UTC
AMD athlon thunderbird 1400@1600, Asus A7V266, kingmax pc2100 ddr.
I get 27.2 fps for the sentinel preset at those settings.
25 flat non overclocked.
Also there doesnt seem to be much difference between thunderbird and XP processors running at the same clock speed. A friend of mine has a XP1700+ (1466 mhz) and when mines at 1466 its pretty much exactly se same.
EL-VIS2
3rd July 2002 12:19 UTC
Real3D Sentinel at 320x240/32BPP Fullscreen
DELL Dimension
PentiumIII/800 with Nvidia GeForce2
--> running at about 29 FPS
dirkdeftly
3rd July 2002 20:36 UTC
Sorry if I sound condescending, but I really don't think you're priorities are straight with upgrading your cpu. You shouldn't be thinking solely of your vis program's performance, instead, you should be considering how your more useful programs will function with a new processor.
But as long as you're asking, the AMD XP 2000+ can calculate Pi to one million digits in three-quarters the time the Pentium P4 can (this seems to me the most useful of the benchmark stats I've seen: pure raw calculation power)
Rovastar
3rd July 2002 20:53 UTC
I think YOU need to get YOUR priorities right. What else do you use your computer for if it is not for visuals?!?!;):D:p:rolleyes:
dirkdeftly
4th July 2002 15:58 UTC
I don't think the word is appropriate - we have kids surfing this forum, remember? :D :D :D :D