Benchmark 2.0

[arobbert]

As has been discussed in several other threads we need a benchmark test with increased complexity to test the newest processors. I feel the most practical way to do this is to modify the existing scene as follows:

1. Increase the Subdivisions of the lightcache to 1000. Using the current settings the calculation is over to quickly.

2. Change the primary bounces to Brute force- Subdivs set to 8. This will greatly increase the load on the processors.

3. Add a 3D text element reading "Benchmark 2.0". This will make the new test easy to identify.

These simple changes will increase the time by roughly four which should allow for a subjective test for the next few years.

I'll volunteer to provide the file in Max 9 and PM percydaman who created the original file for his opinion.

Of course if someone would like to offer a new scene that would be great!

Maybe chaos could provide an official hosting location for the file. Vlado?

We could post the results on this thread.

[meanadam]

We've been using the scene from http://www.3dspeedmachine.com/?page=3 takes 17-40 mins on our machines.

Could we also start a sticky for benchmarking? Maybe a database on vray.info?

Here are some of the results for this above scene (we have more):

i7 930 2.8 ghz (stock), XP 32 bit, 4gb ram = 25:43
i7 930 4.2 ghz (oc), XP 32 bit, 4gb ram = 17:10
QX6700 3157 mhz (oc), XP 32 bit, 4gb ram = 34:25
AMD 620 3600 mhz (oc), XP 32 bit, 4gb ram = 36:53

working on this discrepancy between intel and AMD.... we think it has something to do with the fur.....

Does anybody have a AMD 2.8 Thuban? we are interested in seeing a 6 core AMD result.

[arobbert]

We've been using the scene from http://www.3dspeedmachine.com/?page=3 takes 17-40 mins on our machines.

Could we also start a sticky for benchmarking? Maybe a database on vray.info?

Here are some of the results for this above scene (we have more):

i7 930 2.8 ghz (stock), XP 32 bit, 4gb ram = 25:43
i7 930 4.2 ghz (oc), XP 32 bit, 4gb ram = 17:10
QX6700 3157 mhz (oc), XP 32 bit, 4gb ram = 34:25
AMD 620 3600 mhz (oc), XP 32 bit, 4gb ram = 36:53

working on this discrepancy between intel and AMD.... we think it has something to do with the fur.....

Does anybody have a AMD 2.8 Thuban? we are interested in seeing a 6 core AMD result.

We could just start using that benchmark. My only complaint is that it doesn't have many of the newest processors using Vray. It's set up nice and all the works done with a database and everything.

There's also Frybench which is very up to date but not Vray.
http://randomcontrol.com/index.php?o...=140&benchid=1

[jujubee]

3dspeedmachine bench is ugly.

[arobbert]

Yeah, I agree but they do have a database setup with a Vray test. Not much participation compared to Frybench though.

Nobody appears to be interested in this here.

[Morne]

yes I don't bother with it since there are only handfull of people to compare against

[meanadam]

Who cares what it looks like? It's only a test. Just run it.

[jujubee]

Just saying that it's ugly. And if it's ugly, then the chances are the person who set it up doesn't know what they're doing imo.

[peterguthrie]

Wasn't Vlado going to do some kind of (RT based?) standalone benchmark? hint hint

Maybe as arobbert suggests it could just be based on the existing one

[animator75]

Did a benchmark ever come out from this thread?

[arobbert]

No. I think everyone's focused on RT and video cards.

[Dariusz Makowski (Dadal)]

Just saying that it's ugly. And if it's ugly, then the chances are the person who set it up doesn't know what they're doing imo.

Thats not true ;s just because it doesn't looks amazing it doesn't mean that the guy doesn't know what to do... Maybe hes CGI technician that work with artists that tell him how to set light and not CGI Artist...

I didn't see the render so dunno but well rotfl spamed +1 -.-