understanding SPP

Come here for help & support.
Post Reply
6 posts • Page 1 of 1
User avatar
Pibuz
1st Place 100
Posts: 2646
Joined: Tue Dec 11, 2007 7:58 am
Location: Padua, Italy
3D Software: SketchUp

understanding SPP

Post by Pibuz » Wed Mar 31, 2010 5:59 am

Hi guys.

I just finished my lessons about Indigo-SkIndigo, and in one of the final tests one thing ame up.
I couldn't explain why, so here I am with my question.

The exact same scene was given to all the students, I report here a comparison between two results to show the difference and point out the question.

Like I said, the base scene is the same (during the lessons we modeled a detailed Barcelona Pavillion), exceptede a little variation for the water plane, which was randomly created with a SU plugin.

The first render was computed on a:
core2duo T5800 2GHz ram 2GB
and reached 312 s/px before the desired halt time (30mins)

The second render was computed on a quite older:
pentium centrino 2GHz ram 1GB
and reached 365 s/px in the same amount of time!

This sounded a little weird, if I might say. I couldn' explain why a slower processor could reach a higher SPP value. I've always thought to the SPP value as an objective parameter to determine the quality of renders in terms of noise: an image with 15000 s/px is more noise-free if compared to a 4000 s/px.

I know that sometimes different mats and geometry can make the real difference, but in this case the scene was exactly the same: same model, same scene setup, same materials (all diffuse), same lights, same environment...

Another strange thing is that the same scene looks more noise-free in the image with less s/px.

Can someone explain why this happens?
Thank you a lot!

Image

Image

User avatar
Meelis
Posts: 383
Joined: Sat Mar 14, 2009 11:29 pm
Location: Estonia

Re: understanding SPP

Post by Meelis » Wed Mar 31, 2010 6:18 am

Maybe the faster cpu rendered on 1 core.
Or 1 core was under some nasty program like nero scout (if not disabled).

User avatar
Pibuz
1st Place 100
Posts: 2646
Joined: Tue Dec 11, 2007 7:58 am
Location: Padua, Italy
3D Software: SketchUp

Re: understanding SPP

Post by Pibuz » Wed Mar 31, 2010 10:41 am

Checked while running.
No strange or leech programs running in the background, and the faster notebook was actually using the 2 cores, while the centrino the only core it has.. :twisted:

Jeff
Posts: 112
Joined: Fri Aug 28, 2009 1:28 pm

Re: understanding SPP

Post by Jeff » Wed Mar 31, 2010 4:28 pm

I did not encounter this when I did the renders for the GH house archviz challenge. (Or I may not have noticed it) I used a core2quad @2.3 ghz w/ 3gb ram and a core2duo @2.0 w/ 1gb ram to render several images which I merged later. Renders on the core2duo took twice as long to finish as those from the core2quad. I set the rendering to stop at 2000spp.

User avatar
Meelis
Posts: 383
Joined: Sat Mar 14, 2009 11:29 pm
Location: Estonia

Re: understanding SPP

Post by Meelis » Thu Apr 01, 2010 2:00 am

Can it be virus, or the cpu really is faster than the dual core, cant find benchmarks of centrino product line.
There are new Centrino 2 cpu's http://www.intel.com/products/centrino/

If possible, run another test with some benchmark
http://www.passmark.com/products/pt.htm
Or simple file compression benchmark, multi photo batch processing benchmark.

User avatar
CTZn
Posts: 7240
Joined: Thu Nov 16, 2006 4:34 pm
Location: Paris, France

Re: understanding SPP

Post by CTZn » Thu Apr 15, 2010 10:32 am

The noiser scene seems to have a surface leak for the water, light might enter there and bounce below the water surface a lot (for no visible result but ressources waste).

Even if it's not a geometry leak, it's a detail to dig.

edit: make sure that both scenes use the same mesh and displacement settings for the water surface.
obsolete asset

Post Reply
6 posts • Page 1 of 1

Who is online

Users browsing this forum: No registered users and 46 guests