Seems that the upcoming NVIDIA Kepler cards will be the way to go regarding V-Ray RT rendering.
NVIDIA Kepler GK104 gets 1536 CUDA cores
http://www.nordichardware.com/news/71-graphics/45257-nvidia-kepler-gk104-gets-1536-cuda-cores.html
If I remember correctly it is three times the amount of current GTX 580 cards.
Can’t wait 
Best regards,
Vlado
Yup, throw two of these puppies in and you have the equivalent of six GTX-580s! …And the GPU clock freq starts at 950MHz!
The future looks better all the time…
-Alan
I hope to buy a something of it has half core and eats half power consumption.
I think this is also supposed to be the weakest of the Kepler cards… the bigger models are coming out later in the year.
Best regards,
Vlado
I guess now its a good time to start whinning about render passes support for GPU ! =D
Yeah, they’re saying it will be the entry level card..sheesh!
-Alan
They are already planning the dual core card 
If they will give us 1gb model and ask £6k for 6gb model I go buy shotgun and go after nvidia HQ ! >.<
Yes, the problem, again, is the RAM. This is a game card, so, 2 or 3 GB is enough to play with Crisis4 or UT.
With 2 Gb RAM we can only render some nice, but not heavy, scene. In practice, only modern interiors and objects design.
It’s ok if your job is “only” modern catalogue (and I’ve friend, in two weeks, with different RT engine) can create a 20 or 30 images 5000px noiseless with 6 or 8 580
IMO VRay-Rt GPU is a little behind the others. And, I’ve test here that, with the same noise lever, VRay-RT GPU is 1/2 slow.
I don’t know if it’s the difference from CUDA and OpenCL. But the numbers cannot wrong =)
Anyway, return back about the new GPU, yes, without RAM is not really useful for big interior project or exterior.
We have our library with all the shader usually 4 or 8k, all our model with high polycount etc…
If I will try to think to work with GPU, I think we need to reduce all the texture/models structure etc…
Fight with GPU to see if I’ve enough RAM… mmm… and 6 GB IMO, is not enough! With 12 GB we can start to think in professional mode.
And, if you think that with the new desktop can have not expensive 32 GB RAM ( 4GB x 8 ), with 2 GB on GPU available, I smile!
Where we go with 2 or 4 GB ehehe )))
yeah it will be years to support that, if we look at how they scale things in favor of cores not ram. i
I guess tiled exrs would a solution.
Can you get me specific examples? My own tests show otherwise
Best regards,
Vlado
Vlado - side note how does GPU ram actually works. Is it really that hard to expand it to 20gb or is it just nvidia making loads of $$ by doing it the slow way ?
I don’t know really. There might be technical difficulties to put more RAM on the board.
Best regards,
Vlado
i think the main issues are
a) gpu ram is necessarily very fast as it has to feed all those cores, and hence more expensive than standard ddr dimms.
b) games dont need it yet, and professional uses do, this gives nvidia /amd a nice easy way to differentiate their pro cards, and milk people of their cash.
as to why even the pro cards currently max out at 6 gb or so.. well maybe there are still technical issues running 24 gb of ram at 5 ghz… or maybe its really only us who would use so much (find that hard to believe though. ) not sure tbh.
Hi All.
I love working in the gpu mode ( it’s expansive!! consumes a lot of energy having gpu boxes).
And I really love Vray.
But I must say that working with gpu, it has some problems that competitors have not.
The most obvious problem is the jugged edges. Often they appear and do not go away! I tried 3 gpu competitors engines, and I must say that no one has these problems.
I think because of CUDA…i cannot find any other reason. Also the “look” of the noise is similar for all while in vray is more…how can i say?.. “hard”. it seems that the others have some filter…
The fact is that i have to use 2 kind of software.. one for architectural/animation (vray) and another gpu unbiased for catalogues :(. it’s hard to have 2 material library and convert object…
I hope that vray gpu can achieve the quality /speed of competitors because I think chaos have done great job until now.
PS: in the attachment the example of the jagged edges and (small picture) a quick competitor render.
This is because we don’t have a pixel filter; this is fixed for the next update.
I think because of CUDA…i cannot find any other reason.Nothing of the sort. It’s because we currently don’t have an AA filter on the image.
PS: in the attachment the example of the jagged edges and (small picture) a quick competitor render.Can you post the full resolutions of both renders with the same materials?
Best regards,
Vlado
Vlado, pixel filter will solves the jugged edges problem for both production and RT?
Because if I want a full floating point image in VRay-Producer, with clamp output off, I have always jugged edges…
And I don’t talk if I turn off sub-pixel. x3 x4 rendering time with a lot problem in th every bright area…
It solves it to same extent that it solves it for other renderers…
Best regards,
Vlado
That’s really great! )))
