What does Vray rendering process use: GPU or CPU

[russfogg1]

I'm building a PC which will use Vray as a plugin to Rhino to render still images.

 

I would like to know if I should spend my money on the CPU or the GPU for the fastest rendering times.

 

Any recommendations welcome, but I mainly need to know if the rendering process is carried out by the CPU or GPU (or both).

 

Cheers

[franziska]

Vray 3.6 can use either one and even use them in parallel via CUDA.

Still I am myself not sure if I should first upgrade the CPU or GPU under these circumstances. I lean towards the GPU as the card I am considering Quadro M5000 can live with my existing power and cooling systems of the HP Z800.

[Francisco Penaloza]

Everything is very well explained in the help doc of VRay.

VRay originally was created to work with CPU, and still, all properties and effect will work with the CPU.

lately, with all the GPU frenzy, VRay implemented a GPU version. From the beginning, this version had some limitation, and as up today, still have some limitations but it is mostly compatible with everything VRay, Shaders, Lights etc.

 

Now what you need to learn is that no matter what the video card manufacturer sell you, there is one basic physic constant. MEMORY, no matter how fancy your video card is, if it does not have enough memory for your scene, it won't work!

 

There are some works around and VRay uses some of them when you use both methods, CPU+GPU, but in those case, it is just better to go with CPU because GPU is not supporting much.

 

Now depending on what type of visualization you do, 8 or 16 Gb of memory may be OK.

If you do variate visualization and large projects, you only need to look for CPU.

 

GPU rendering is more popular on Motion graphics and similar because the way project assets are managed, in Architectural Viz, CPU still a solid foundation with no limits, well depending on how many cores you can get, share or rent.

[russfogg1]

Hi, apologies for the late reply...the project went on hold. That's great to know and I've gone with a high spec CPU, decent RAM and a just a reasonable GPU.

 

Should be putting it all together next week and hope to see some big improvements from my laptop.

 

Thanks for your help

 

Everything is very well explained in the help doc of VRay.

VRay originally was created to work with CPU, and still, all properties and effect will work with the CPU.

lately, with all the GPU frenzy, VRay implemented a GPU version. From the beginning, this version had some limitation, and as up today, still have some limitations but it is mostly compatible with everything VRay, Shaders, Lights etc.

 

Now what you need to learn is that no matter what the video card manufacturer sell you, there is one basic physic constant. MEMORY, no matter how fancy your video card is, if it does not have enough memory for your scene, it won't work!

 

There are some works around and VRay uses some of them when you use both methods, CPU+GPU, but in those case, it is just better to go with CPU because GPU is not supporting much.

 

Now depending on what type of visualization you do, 8 or 16 Gb of memory may be OK.

If you do variate visualization and large projects, you only need to look for CPU.

 

GPU rendering is more popular on Motion graphics and similar because the way project assets are managed, in Architectural Viz, CPU still a solid foundation with no limits, well depending on how many cores you can get, share or rent.

[abhisekguha]

Here is what my experience is...

I built a Ryzen 1700x 16 gb ram PC with gtx 1080ti.... but I realized 1080ti is overkill for what I do...

Vray mainly has two modes of operations, Adv and RT. Vray RT do uses CUDA but has many limitation on rendering features and once you are used to with Adv you won't likely switch to RT mode. The Adv mode, on the other hand, is way to go mode for me for its features but it doesn't use even 5% of gpu power! Adv mode renders best and eats up 100% of my 8 Core cpu and brings 16 GB ram to its knees...

 

So my suggestion is, Vray (most of the multimedia softwares for that matter) is yet not optimized to use full potentials of GPU, you need a CPU with as much core and speed as you can afford and 32 GB ram is best.

Edited April 5, 2018 by abhisekguha