OTOY Forums

Ok just got my point of view gtx465, I went from 1.3Msamples/s with my gtx260 to 2.5Msamples/s, quite what expected

I would add that this is the loudest card I have ever heard, I can't even believe they sell such a thing !

2.5 is not bad. keep in mind we still have some optimizations for fermi planned for 2.3 so it could reach 3 Ms.

Radiance

radiance wrote:2.5 is not bad. keep in mind we still have some optimizations for fermi planned for 2.3 so it could reach 3 Ms.

Radiance

good to know

is there a way to reduce the amount of gpu power used by octane ? it gets almost impossible to navigate in windows while octane is running, so beside pausing the render or adding a second graphic card in my pc what can i do ?

radiance wrote:currently only tesla cards with 4GB and i'd wait a bit until fermi based cards are available with more than 4GB ram.
We also need to make some modifications to octane in order to render larger than 4096x4096

Radiance

Tesla cards and supercomputers have up to 24 GB

Another question about software... Can your software use Optix, CompleX and SceniX engines (http://www.nvidia.com/object/optix.html)?

soulandmind wrote:

radiance wrote:currently only tesla cards with 4GB and i'd wait a bit until fermi based cards are available with more than 4GB ram.
We also need to make some modifications to octane in order to render larger than 4096x4096

Radiance

Tesla cards and supercomputers have up to 24 GB

Another question about software... Can your software use Optix, CompleX and SceniX engines (http://www.nvidia.com/object/optix.html)?

No, We don't use optix, or any 3rd party sdk.
That's why octane runs fine on equipotent yet far cheaper Geforce instead of Quadro cards.

Radoamce

using this http://downloads.guru3d.com/ASUS-GTX-46 ... -2561.html I went from 2.5MS/s to 3

I have the beta running the Benchmark and it says 2.69 mega samples.
I have a GT240 with 512 MB Ram.
How good is that?
Thanks.

am i the only one to get less samples/s with the new nvidia 257.21 drivers with gtx470 ?

sam75 wrote:am i the only one to get less samples/s with the new nvidia 257.21 drivers with gtx470 ?

you might want to stick with 197.45WHQL until we upgrade to cuda 3.1.

Radiance

Hello!

Im building a new system right now and all this GPU Render, CUDA and OpenCL is driving me crazy! Im posting this here cause I think people here can help me.

I'll be using this PC for 3D work, mainly 3ds Max, and Octane Render seems a pretty cool Render to use with 3ds Max. I'm not a big professional worker, but I handle some scenes with around 4 million polys, animation etc. My main concern is about GPU. I need a good viewport preview and looking for a good render speed too. It seems pretty obvious that GPU render is the future, but what about OpenCL and CUDA? Looking for the future I bet OpenCL will take the lead, cause it's compatible with more cards, and more cards = more clients = more money. But I don't have any idea about when it would get over CUDA. That's why I want to know about Octane implementation for OpenCL, is there anything already going on? Are you guys having problems with ATI OpenCL? I read some words of Vlado (from Chaos Group) and he said that they were having trouble with it, but still working on.

About the GPU Render performance, what else besides GLFLOP is important? I was looking for 2 PC configs, one of them is a Phenom X6 1090T + FirePro V8800 and the other one is i7 980X + V7800, not sure about which one to get. Those cards would have good performance for GPU rendering? (assuming that they will be compatible with good a GPU Renders as Octane). What would you recommend me to buy for CPU + GPU solution around US$1500? (Remembering that I'm looking for a good viewport and render) Do you know anything about GTX480 performance on 3ds Max viewport?

Guess that's all.. I'll probably rember some things later and come back here

PS: Sorry if I wrote something wrong, Im from Brazil.