mac vs pc rendering speed

Blender's renderer and external renderer export

Moderators: jesterKing, stiv

rednelb
Posts: 0
Joined: Sun Aug 01, 2004 1:34 pm

Post by rednelb »

levon wrote:ill do this benckmark test in a bit, but....

doesnt a benchmark have to test all aspects of the computer/renderer.
like trying different res textures and etc??

ill post my result later.
you are of course right. a good benchmark really should examine the more the better aspects. but i thought myself not every aspect is used equally often, eg radiosity is still very slow and therefore sometimes faked; and as raytracing is both cpu and memory incident it could hopefully sufficiently describe the rendering speed differences between mac and pc.
albeit "there are lies, damn lies, and benchmarks" :).
and thanks for the results.

joeri
Posts: 96
Joined: Fri Jan 10, 2003 6:41 pm
Contact:

Post by joeri »

It might be nice to have a prestige list at blender3d.org, or is there already?
Last edited by joeri on Fri Oct 08, 2004 2:39 pm, edited 1 time in total.

solmax
Posts: 86
Joined: Fri Oct 18, 2002 2:47 am
Contact:

Post by solmax »

here some quick comparisons:

render time, blender internal renderer:

G4 867MHz: 26,57s
G5 2x2GHz: 8,9s (blender uses only ONE processor!)

and here the x86 "competitor"
ATH 600, x86 optimized blender: 28,6s
ATH 600, x86 regular blender: 44,29s

i think there is a lot of potential for optimizing blender for the G4/5 platform, but one of the developers (luke_p) sais it's very difficult. a first step woud be to make at least the renderer multi-threaded so that both processors on a G5 platform can be utilized.

rednelb
Posts: 0
Joined: Sun Aug 01, 2004 1:34 pm

Post by rednelb »

joeri wrote:Jwalton:
It might be nice to have a prestige list at blender3d.org
that's true, that's true. or even better a whole site dedicated to rendering with blender.

rednelb
Posts: 0
Joined: Sun Aug 01, 2004 1:34 pm

Post by rednelb »

solmax wrote:here some quick comparisons:

render time, blender internal renderer:

G4 867MHz: 26,57s
G5 2x2GHz: 8,9s (blender uses only ONE processor!)

and here the x86 "competitor"
ATH 600, x86 optimized blender: 28,6s
ATH 600, x86 regular blender: 44,29s

i think there is a lot of potential for optimizing blender for the G4/5 platform, but one of the developers (luke_p) sais it's very difficult. a first step woud be to make at least the renderer multi-threaded so that both processors on a G5 platform can be utilized.
thank you! and please, could you also provide the settings you used since i assume that the settings of my benchmark were not very valid for internal renderer. or even better, if you would have the relish, try it with yafray?

ed catmull (pixar + subdivision sufaces, i think) said "that the g5 is the fastest desktop in the world" in pixar's renderman benchmark. is he right?

and finally, for luke_p and anyone interested:
Optimizing for the Power Mac G5 http://developer.apple.com/performance/ ... ation.html
G5 Performance Programming http://developer.apple.com/hardware/ve/g5.html
PowerPC G5 Performance Primer http://developer.apple.com/technotes/tn/tn2087.html
Tuning for G5 http://developer.apple.com/technotes/tn/tn2086.html

bertram
Posts: 0
Joined: Wed Oct 16, 2002 12:03 am

Post by bertram »

I don't know wheter this test is representative or not, but there's a built-in benchmarking feature inside blender.
Just select <help> from the header menu and there's "benchmark" in the 2nd level under <system>.
At my sys the result sounds like: 429,476684 FPS (2148 iterations)

Does anyone exactly knows what this test ist benchmarking?

leope
Posts: 0
Joined: Mon Apr 05, 2004 1:26 pm

Post by leope »

1:36:30
athlon 3800+, 1GB 2326
yafray 0.0.7 compiled on fedora core 2 64bit[/b]

solmax
Posts: 86
Joined: Fri Oct 18, 2002 2:47 am
Contact:

Post by solmax »

bertram wrote:I don't know wheter this test is representative or not, but there's a built-in benchmarking feature inside blender.
Just select <help> from the header menu and there's "benchmark" in the 2nd level under <system>.
At my sys the result sounds like: 429,476684 FPS (2148 iterations)

Does anyone exactly knows what this test ist benchmarking?
this is the viewport (OpenGL) performance.

Marty_D
Posts: 0
Joined: Mon Jun 09, 2003 11:16 pm

3:56

Post by Marty_D »

rednelb wrote:hi all,
hi rednelb, here you go:
render time - 3 min 56 sec
machine - Pentium 4, 2.4 ghz, 1.5 Gb ram (no MT)
soft - Blender 2.34 official, Yafray 0.0.7 [edit] Windows XP sp1 [/edit]
--
Last edited by Marty_D on Mon Feb 21, 2005 7:36 pm, edited 1 time in total.

joeri
Posts: 96
Joined: Fri Jan 10, 2003 6:41 pm
Contact:

Post by joeri »

Wait, what scene are you rendering?
I can imagine you want to know what machine is the bestest, but I think we should stick to the internal renderer, or is that to strict?

JWalton
Posts: 58
Joined: Sun Oct 13, 2002 7:39 pm
Contact:

Post by JWalton »

i think you can take the 'compatible graphics card' thread as a failed attempt to mine any useful information. there hasn't been a good solid reliable success at blenchmarking anything since the NaN days.

Monkeyboi
Posts: 251
Joined: Tue Nov 26, 2002 1:24 pm
Location: Copenhagen, Denmark
Contact:

Post by Monkeyboi »

If you use Yafray for testing, what you are really testing is, well, Yafray. Please test with Blender internal.
I can give you Mac/PC comparisons when I get my new iMac through the door :wink:

rednelb
Posts: 0
Joined: Sun Aug 01, 2004 1:34 pm

Post by rednelb »

hi everybody,
since a few people put in question the decision of using yafray, i owe you a little clarification on this.
first of all, in context of rendering speed comparision there is no difference what renderer you use as long as you employ the same one for rest of tests.
so why did i chose the yafray then?
as hard as i try to imagine some bulletproof arguments, those fail short after the really simple one - i like yafray more.
now what? since vastly majority of us employed yafray for this benchmark already and the scene is prepared here and now i suggest we keep using yafray.
any comments? please keep in mind this has not to be yafray vs internal renderer.
thanks for every reply so far and keep posting!

joeri
Posts: 96
Joined: Fri Jan 10, 2003 6:41 pm
Contact:

Post by joeri »

I'm a bit dislectic and it took me a while to find this scene.

Since I don't have yafray installed I can't help with the benckmark.

Frankly I think benchmarking yafray is a yafray forum thing, we might as well be benchmarking maya 'cause blender can export to obj.
I mean it in a friendly way. I can see how blender is more connected to yafray, but I'd rather see an out-of-the-box compare, because the response could be wider. And it would be a "How fast is blender" compare.

Now what? you ask, well keep them coming, and maybe you can summerize the results in a nice table when this threat turns dead.

LetterRip
Posts: 0
Joined: Thu Mar 25, 2004 7:03 am

Post by LetterRip »

Why don't we set up a test page at the Wiki?

Have a selection of different scenes and what particular effects should be used

The a list that contains both the computer and renderer

I can set it at the wiki if there is interest, but someone else will have to design some test scenes and such.

I'm already in the process of creating a massive feature comparison table for the assorted graphics programs, so this would be along similar lines.

LetterRip

Post Reply