Hello guys,
I’ve just made some test between SP4 (max 2009) and SP3a (max 2008), 64bits version.
Workstation : Core i7 920@3000ghz 12Gb hyper-threading ON
Here are the results :
2008 Sp3a : 10min25sec.2009 SP4 : 11min18sec.
Any idea ? SP4 shouldn’t be faster or I missed something ?
You can play with max scene (2008 only, we don’t use max9 anymore), see attachments.
Last think I will test is 2008 + SP4TestSP4.zip (22 KB)
New try :
2008 64bits SP4 : 10min38sec.
Sorry, my bad… maybe I was expecting more (or less rendertime…)
But SP4 isn’t faster as SP3a, I was hoping that multi-core improvement will speed up rendering.
Best regards…
PS: Maybe it’s not in the right forum’s section. It’s not a problem. Maybe in General section ?
my guess is that it will only speedup in specific renderings or settings
That’s why I try : GI + Glossy refl + Displacement + Refraction
would be nice to know ?
Best regards…
Neilg
October 29, 2009, 9:21am
5
Your scene may be too small - it may be when it hits the wall and memory gets tight that the differences start to come out.
from what I can tell the main difference is in the memory handling of larger scenes.
I’ll try with a bigger one, with more geometries to see any difference.
Hope to post one quickly.
Best regards…
Hello here a new test…
More geometries (close to 4Gb)
Max 2008 64bits : 17min43sec.
Nothing different with a more complex scene.
vlado
October 29, 2009, 4:12pm
9
The difference only shows up in specific cases; for other scenes, there shouldn’t be any significant changes in render times.
Best regards,
do we have ourselves here finaly a new cool benchmark scene?
Codi
October 29, 2009, 5:48pm
11
(Still trying to download SP4)
The improvement is for scenes with a lot of dynamic geometry? (VrayDisplacement, VRayFur, …)
n6
October 29, 2009, 8:44pm
12
on a 1 200 000 poly scene that was rendering in 6min44s with sp3 it renders in only 6min22s with sp4.
Test was done on dual XEON E5520 so 16 core system, 12Gb memory.
For me, sorry, but SP4 is faster than sp3.
n6:
on a 1 200 000 poly scene that was rendering in 6min44s with sp3 it renders in only 6min22s with sp4.
Test was done on dual XEON E5520 so 16 core system, 12Gb memory.
For me, sorry, but SP4 is faster than sp3.
I’m no saying that SP4 is useless :). But as I say in previews post, I was hoping faster result, that’s I’m wondering what part of the rendering take advantage of that optimization.
My current test :
And I see no difference between SP3a and SP4. Maybe a more complex scene, with fur, caustics, bitmap texture.
I’m thinking about providing a big benchmark with all this feature in it…
Best regards…
