Thea Render Forum

As far as I know they changed the way they count CUDA power. It is not the same core unit, so it is not comparable between 5.. and 6.. series.

saurus wrote:As far as I know they changed the way they count CUDA power. It is not the same core unit, so it is not comparable between 5.. and 6.. series.

that's interresting, at first I thought there was some error during the test, but there are many similar results

Kepler(600/700) vs Fermi (older 400/500) is it because the cuda support still doesn't take full advantage of Kepler?

6xx and 7xx chips were crippled in order to sell more Quadro cards. The 5xx chips were performing nearly as well as the Quadros, so they changed the rules. But, this really only applies to certain operations. The performance benefit shows up most in screen draw, not GPU processing like rendering. Being able to handle huge datasets of points and lines in a viewport is what a Quadro excels at. Their render performance, not so much.

You'll see that the GFl of a GTX680 is significantly higher than a K5000; having the same number of CUDA cores.

The official pdf says that gtx 580 will make 410 s/px in caustics (http://thearender.com/cms/images/editio ... chmark.pdf). I have this card and on high priority I get about 700 s/px.

It also says that Titan makes 800 s/px , so is it right to assume that it will make 1400 s/px on high priority. Can anyone test it out? It can be 780ti instead cause they are similar.

choppedporks wrote:The official pdf says that gtx 580 will make 410 s/px in caustics (http://thearender.com/cms/images/editio ... chmark.pdf). I have this card and on high priority I get about 700 s/px.

It also says that Titan makes 800 s/px , so is it right to assume that it will make 1400 s/px on high priority. Can anyone test it out? It can be 780ti instead cause they are similar.

I wanted to keep the results not exaggerating (like others... coming with statements such as the gpu is running 50x faster than the cpu...). We ran most of them on graphic cards with Low priority (which makes a big performance on the graphic card but it is the safest priority which also leaves your machine responding) plus there were additional optimizations for Presto on the way despite adding also more features. On top of that, the Xeon processors and i7-6core benchmarks have been also improved (with 1.3.05).

Also, please keep in mind that the overall system may affect the graphic card performance. For example, at some points the results need to be transferred back to CPU for display/autosave; I wouldn't say this is a stumbling block but you should expect something better keeping the same card but improving the rest of the system (not much though).

To answer your question, the Titans should be slightly better but not as high as 1400 s/p.

Best wishes

Thanks Giannis,
I know you put the safe numbers in .pdf and that's ok. I just wanted to know what is the full power of hardware because I make a lot of renders and time is an important factor for me.

Cheers

choppedporks wrote:Thanks Giannis,
I know you put the safe numbers in .pdf and that's ok. I just wanted to know what is the full power of hardware because I make a lot of renders and time is an important factor for me.

Sure, I get it

Generally speaking, the benchmark should be treated as an indication; there will be ups & downs in performance with every different revision, every different system, scene & settings. But what you can do, is to run the benchmark scenes on your system and see where your system stands within all the other systems that we have published.

Best wishes

Hi everyone,
I was pleasantly surprised by the performance of my laptop in these benchmark.
It seems that setting my gtx880m to high priority makes it perform better than a titan with low priority

Here are the results:

Caustics scene

Opening file "C:\ProgramData\Thea Render/Scenes/Benchmark/Presto/Caustics/Caustics.scn.thea"
Finished in 0 seconds!
Device #0: GeForce GTX 880M
Building Environment... done. (0.16 seconds)
- Polys: 13520, Objects: 5, Parametric: 0
- Moving: 0, Displaced: 0, Clipped: 0
- Instances: 0, Portals: 0
- Nodes: 4199, Leaves: 4200, Cache Level: Normal
Building Environment... done. (0.15 seconds)
- Polys: 13520, Objects: 5, Parametric: 0
- Moving: 0, Displaced: 0, Clipped: 0
- Instances: 0, Portals: 0
- Nodes: 1336, Leaves: 4009, Cache Level: Normal
Device #1: Intel(R) Core(TM) i7-4810MQ CPU @ 2.80GHz
Device #0: memory 330/8192 Mb
Device #1: memory 1739/16300 Mb
Device #0: 509x4 s/p
Device #1: 239x4 s/p
Finished in 5 minutes and 2 seconds!

Instancing scene

Opening file "C:\ProgramData\Thea Render/Scenes/Benchmark/Presto/Instancing/Instancing.scn.thea"
Finished in 0 seconds!
Device #0: GeForce GTX 880M
Building Environment... Note: activating instancing mode for saving memory.
done. (0.671 seconds)
- Polys: 138667997609, Objects: 1063033, Parametric: 0
- Moving: 0, Displaced: 0, Clipped: 531516
- Instances: 1063028, Portals: 0
- Nodes: 646652, Leaves: 646657, Cache Level: Normal
Building Environment... Note: activating instancing mode for saving memory.
done. (0.672 seconds)
- Polys: 138667997609, Objects: 1063033, Parametric: 0
- Moving: 0, Displaced: 0, Clipped: 531516
- Instances: 1063028, Portals: 0
- Nodes: 167262, Leaves: 501791, Cache Level: Normal
Device #1: Intel(R) Core(TM) i7-4810MQ CPU @ 2.80GHz
Device #0: memory 692/8192 Mb
Device #1: memory 3101/16300 Mb
Device #1: 57 s/p
Device #0: 62 s/p
Finished in 5 minutes and 2 seconds!

Meccano plane scene

Opening file "C:\ProgramData\Thea Render/Scenes/Benchmark/Presto/MeccanoPlane/MeccanoPlane.scn.thea"
Finished in 0 seconds!
Device #0: GeForce GTX 880M
Building Environment... done. (0.140 seconds)
- Polys: 503556, Objects: 91, Parametric: 0
- Moving: 0, Displaced: 0, Clipped: 0
- Instances: 0, Portals: 0
- Nodes: 158062, Leaves: 158063, Cache Level: Normal
Building Environment... done. (0.110 seconds)
- Polys: 503556, Objects: 91, Parametric: 0
- Moving: 0, Displaced: 0, Clipped: 0
- Instances: 0, Portals: 0
- Nodes: 40851, Leaves: 122554, Cache Level: Normal
Device #1: Intel(R) Core(TM) i7-4810MQ CPU @ 2.80GHz
Device #0: memory 387/8192 Mb
Device #1: memory 2092/16300 Mb
Device #0: 365x4 s/p
Device #1: 208x4 s/p
Finished in 5 minutes and 1 seconds!

SSS gears scene

Opening file "C:\ProgramData\Thea Render/Scenes/Benchmark/Presto/SSSGears/SSSGears.scn.thea"
Finished in 0 seconds!
Device #0: GeForce GTX 880M
Building Environment... done. (0.46 seconds)
- Polys: 48218, Objects: 23, Parametric: 0
- Moving: 0, Displaced: 0, Clipped: 0
- Instances: 0, Portals: 0
- Nodes: 13736, Leaves: 13725, Cache Level: Normal
Building Environment... done. (0.16 seconds)
- Polys: 48218, Objects: 23, Parametric: 0
- Moving: 0, Displaced: 0, Clipped: 0
- Instances: 0, Portals: 0
- Nodes: 4211, Leaves: 12621, Cache Level: Normal
Device #1: Intel(R) Core(TM) i7-4810MQ CPU @ 2.80GHz
Device #1: memory 1867/16300 Mb
Device #0: memory 337/8192 Mb
Device #0: 101x4 s/p
Device #1: 99x4 s/p
Finished in 5 minutes and 4 seconds!

Finally got a decent machine.

Opening file "C:\ProgramData\Thea Render/Scenes/Benchmark/Presto/Caustics/Caustics.scn.thea"
Finished in 0 seconds!
Device #0: Quadro K4200
Building Environment... done. (0.16 seconds)
- Polys: 13520, Objects: 5, Parametric: 0
- Moving: 0, Displaced: 0, Clipped: 0
- Instances: 0, Portals: 0
- Nodes: 4199, Leaves: 4200, Cache Level: Normal
Building Environment... done. (0.16 seconds)
- Polys: 13520, Objects: 5, Parametric: 0
- Moving: 0, Displaced: 0, Clipped: 0
- Instances: 0, Portals: 0
- Nodes: 1336, Leaves: 4009, Cache Level: Normal
Device #1: Intel(R) Xeon(R) CPU E5-2630 v3 @ 2.40GHz
Device #1: memory 4707/32691 Mb
Device #0: memory 301/4096 Mb
Device #1: 827x4 s/p
Device #0: 376x4 s/p
Finished in 5 minutes and 1 seconds!

Opening file "C:\ProgramData\Thea Render/Scenes/Benchmark/Presto/Instancing/Instancing.scn.thea"
Finished in 0 seconds!
Device #0: Quadro K4200
Building Environment... Note: activating instancing mode for saving memory.
done. (0.889 seconds)
- Polys: 138667997609, Objects: 1063033, Parametric: 0
- Moving: 0, Displaced: 0, Clipped: 531516
- Instances: 1063028, Portals: 0
- Nodes: 646652, Leaves: 646657, Cache Level: Normal
Building Environment... Note: activating instancing mode for saving memory.
done. (0.858 seconds)
- Polys: 138667997609, Objects: 1063033, Parametric: 0
- Moving: 0, Displaced: 0, Clipped: 531516
- Instances: 1063028, Portals: 0
- Nodes: 167262, Leaves: 501791, Cache Level: Normal
Device #1: Intel(R) Xeon(R) CPU E5-2630 v3 @ 2.40GHz
Device #0: memory 733/4096 Mb
Device #1: memory 6081/32691 Mb
Device #1: 217 s/p
Device #0: 52 s/p
Finished in 5 minutes and 4 seconds!

Opening file "C:\ProgramData\Thea Render/Scenes/Benchmark/Presto/MeccanoPlane/MeccanoPlane.scn.thea"
Finished in 0 seconds!
Device #0: Quadro K4200
Building Environment... done. (0.203 seconds)
- Polys: 503556, Objects: 91, Parametric: 0
- Moving: 0, Displaced: 0, Clipped: 0
- Instances: 0, Portals: 0
- Nodes: 158062, Leaves: 158063, Cache Level: Normal
Building Environment... done. (0.156 seconds)
- Polys: 503556, Objects: 91, Parametric: 0
- Moving: 0, Displaced: 0, Clipped: 0
- Instances: 0, Portals: 0
- Nodes: 40851, Leaves: 122554, Cache Level: Normal
Device #1: Intel(R) Xeon(R) CPU E5-2630 v3 @ 2.40GHz
Device #0: memory 443/4096 Mb
Device #1: memory 5051/32691 Mb
Device #0: 305x4 s/p
Device #1: 726x4 s/p
Finished in 5 minutes and 0 seconds!

Opening file "C:\ProgramData\Thea Render/Scenes/Benchmark/Presto/SSSGears/SSSGears.scn.thea"
Finished in 0 seconds!
Device #0: Quadro K4200
Building Environment... done. (0.31 seconds)
- Polys: 48218, Objects: 23, Parametric: 0
- Moving: 0, Displaced: 0, Clipped: 0
- Instances: 0, Portals: 0
- Nodes: 13736, Leaves: 13725, Cache Level: Normal
Building Environment... done. (0.31 seconds)
- Polys: 48218, Objects: 23, Parametric: 0
- Moving: 0, Displaced: 0, Clipped: 0
- Instances: 0, Portals: 0
- Nodes: 4211, Leaves: 12621, Cache Level: Normal
Device #1: Intel(R) Xeon(R) CPU E5-2630 v3 @ 2.40GHz
Device #0: memory 373/4096 Mb
Device #1: memory 4703/32691 Mb
Device #1: 338x4 s/p
Device #0: 87x4 s/p
Finished in 5 minutes and 3 seconds!