OTOY Forums

Hi,

I've a little question. Why when Octane run, mt Windows XP64bit is so slow? Also play with octane is slow. Move the node in the graph is slow. Is there an option to change the priority in the GPU? Like for the CPU on windows? Or another solution?

Add extra INFO:

9800 GTX
Cuda Version 3.00
Cuda Runtime 2.03
Capability Mode 1.1

Maybe I'm wrong but I think this isn't possible.
Means you would need a secondary GPU. One that runs Octane and one that runs the viewscreen.

Would be great indeed if there is a way to "adjust" the calculation priority that you can still work in other windows

No, this is a limitation of GPU computing.
GPU's don't have multitasking operating systems on them and can only run 1 app at the same time, either a render engine or your windows user interface.

rendering with a 2nd GPU removes this.
it's not octane specific, but all GPGPU applications do it.

I'm hoping nvidia finds a solution for it.

Radiance

Thanks radiance
I will buy another cheap VGA

radiance wrote:No, this is a limitation of GPU computing.
GPU's don't have multitasking operating systems on them and can only run 1 app at the same time, either a render engine or your windows user interface.

rendering with a 2nd GPU removes this.
it's not octane specific, but all GPGPU applications do it.

I'm hoping nvidia finds a solution for it.

Radiance

I am new to GPU computing.
So does that mean that if I add a second video card to my system, it will speed things up?
Would i be able to select which card to do the work?
I am asking because my current card is GT8800 which has only 512Mb ram.
If I add a second card, say a 1Gb card, would I be able to selectively pass the rendering work to the 1Gb card and while it is rendering,
continue to work on other tasks with the 512Gb card?

yes, in the device manager panel in octane you can choose the GPU to use.
you can only change it after starting octane, before you start the render engine by clicking on a meshnode.

if you have 2 cuda capable gpus, it will automatically use the 2nd one by default.

use the fastest one as the 2nd, and the slower one as the first connected to your display.

Radiance

I'm a bit ignorant when it comes to hardware connections.

Radiance, when you say 1st and 2nd, does it mean you connect the first card (slower 512MB) to the 16x PCI-e slot on the motherboard and the second card (faster 1gig) to the 8x PCI-e slot?

Also, would the smaller 512MB card not limit the ram to Octane to 512MB?

Thanks,

jershaun wrote:I'm a bit ignorant when it comes to hardware connections.

Radiance, when you say 1st and 2nd, does it mean you connect the first card (slower 512MB) to the 16x PCI-e slot on the motherboard and the second card (faster 1gig) to the 8x PCI-e slot?

Also, would the smaller 512MB card not limit the ram to Octane to 512MB?

Thanks,

connect the fastest card to the fastest slot. (this is now secondary GPU, for octane)
connect the display to the other card in the slower slot. (this is now primary GPU)

depending on which slot you placed them in (on MB it will say PCIEx16 1, and PCIEx8 2 or something), the order will change in the octane device manager.

Radiance