Printable View
Just played around with the Demo Release of Cinema4d. I wanted to figure out performance differences between 10.0 and 11.0 Release for a K10. So i took the model from cinbench and rendered it in Cinema4d 11.0 Demo Release and noted the time the picture was done. Ok pictures says more then 1000 words ;)
http://www.abload.de/img/c4dpzy.jpg
Nice work Maxon :clap:. Would be nice to see increases of a C2Q :)
ftp://ftp.maxon.net/pub/r11/CINEMA4DR11.zip
whut
^
Yes, I'm that stoked. C4D was my main app, but too late- I'm moving to greener pastures. The renders are still ugly and last-gen compared to Finalrender, vRay, Renderman, Mental or Maxwell.
Oh man,that's some optimization job right there.It would be nice to see the increase on all platforms.
Vector SSE??Quote:
Originally Posted by informal
I would like to think they finally did that,but i honestly don't know.Using scalar SSE when there is vector which is like miles faster always made me wonder wth are they doing there...
More like they fixed the ****ing borked Global Illumination.
That was AWFUL in R8.5, R9, R10, R10.5... when other software packages already had tons of improvments in rendering.
Any Core2 users willing to run some tests?
I can try, but I only ever ran Cinebench and never really used the program ;)
If I get it to work I can compare times with CB10 on the Q6600 as well as my Octacore (rigs in sig)
with Core 2 Duo E8600 @ 4Ghz
Cinebench R10 = 1min 45sec
Cinema 4D 11 = 45sec
Cinebench R10 64 bit : 1 min 01 sec
Cinema 4D R11 64 bit : 28 sec
Yet another benchmark :)
Quote:
Originally Posted by Chri$ch
wow, nice improvement :shocked:Quote:
Originally Posted by Toysoldier
I tryed to implement the Cinema4d 11 engine to Cinebench and it works :D. If you want the modded Cinbench10 just -> PM.
Isn't a rendering time improvement in the default render pretty much useless in real life applications?
I'm a max user, and I have never ever done any final renders with the default rendered. I almost always use vray or mental ray, even when baking textures.
Do you never prerender modell objects just for checking out the picture or scene?Quote:
Originally Posted by SafeFire
An example of the "modded" Cinebench10-11 using your files..Quote:
Originally Posted by Boschwanza
Damn it's slow..:rofl:
http://img50.imageshack.us/img50/7816/cb10113800rm6.jpg
Hehe, thats what Intel did on IDF with its 45.000 point on Nehalem ;)
If your REALLY sure of that then I have to send this to fugger..:rofl:Quote:
Originally Posted by Boschwanza
I got the same results movieman, Check it.
http://i31.photobucket.com/albums/c3...Untitled11.png
Its just a bit logical . Cinema4d 11 Render engine will give you a boost up to 2.5x.
Nehalem 2.93 scores about 14k points with Cinebench 10 32Bit .. 3.2 will be around 17.000 * 2.5 makes 42.500 .. pretty close isnt it ? :D
I'd just like to know for sure if this is what they did because if that is so then my Harpertown at 3800 just destroyed their Nehalem and that doesn't make sense. They would bring their "A" game to IDF and no old fart like me is going to top them by 10,000 points..Quote:
Originally Posted by Boschwanza
( as much as I'd like to beleive it!):rofl:
UNLESS their 45,000 score was with a single socket 4 core...HMMMMMMM...
Thats what i thaught :shrug: at least what has be reported here in germanyQuote:
Originally Posted by Movieman
Then it makes sense with 4 cores and if you look at warboys score at 3600mhz of 31,000 their 45,000 and my 55,000 make sense..Quote:
Originally Posted by Boschwanza
Now I wonder what their 8 core Gainstown at 3200 will do? :sofa:
if they modded cinbench they are just cheating numbers, cause average Joe thinks Nehalem will be 3 times faster then a QX9770 :confused:.Quote:
Originally Posted by Movieman
Anyway 45000 is still very impressiv even with Cinema4d 11.0.
I had seen earlier that I had to run the harpers at 3600 to match a gainstown at 3000 and he was only on single channel memory and beta bios.Quote:
Originally Posted by Boschwanza
I think Sampsa said that he saw the number(45K) and it was done on single socket "Nehalem 3.2Ghz overclocking system" ,3.2Ghz meaning it's the starting/default clock.I would expect that the score of 45K is with default or slightly OCed chip,since it fits perfectly with numbers Boschwanza provided in his post.
Quote:
Originally Posted by Boschwanza
+400MB so bigg :s
Well, So Intel cheated at Cinebench, It's not the first time that a company cheated at a benchmark.
downloading .. now i will see how think improved in the new version :-)
Yeah, of course I do non-final renders, tons of them. I just use the rendered I would use for the final render, what good is checking out a scene in the default renderer, only to find out that's it not fine when using Vray or Mental Ray?Quote:
Originally Posted by Boschwanza