Puget Systems print logo

https://www.pugetsystems.com

Read this article at https://www.pugetsystems.com/guides/1807
Article Thumbnail

Best Workstation PC for V-Ray Next GPU (Spring 2020)

Written on June 15, 2020 by William George
Share:

Introduction

Here at Puget Systems, specifically in the Labs department, most of the content that we write falls into one of two categories: either long, fairly in-depth articles looking at the performance of various PC components in a given application or recommended system pages, with multiple configuration options aimed at different budgets and lots of supporting data drawn from the aforementioned articles. For some readers all of that is information overload, though, and they just want a quick question answered: what is the best computer for my needs? We’re going to try answering that question more directly in a new series of short articles, like this one.

Today we are looking at V-Ray Next from Chaos Group, and specifically the GPU based side of this rendering engine. As applications go, it is pretty simple in its needs: video cards - the more and the faster the better, with lots of VRAM. These recommendations will also apply to older versions of V-Ray, as long as you are using the GPU processing side of things (we have another article for the CPU side).

Screenshot of V-Ray Next Benchmark

Screenshot of V-Ray Next Benchmark

Good PC Configuration for V-Ray Next GPU

We'll start off with a good, solid workstation design for V-Ray Next - hopefully without breaking the bank. Compared to other GPU-based rendering engines we have tested, V-Ray Next benefits far less from higher-end video cards (at least with the same amount of VRAM) but it does scale very well with multiple cards. To start off, then, we have a tower that supports three GPUs with AMD's Threadripper processor platform. This is a bit overkill on the CPU side, but allows us to get up to three video cards without compromising on PCI-Express lanes; the more consumer-oriented platforms either max out at two GPUs or start restricting PCI-E lanes to 4x speeds in order to sneak in a third card. V-Ray Next does also allow the CPU to be used alongside the video cards, in a sort of GPU-emulation, but the performance gained there is not nearly as cost-effective as having more real video cards. To round the system out we have 64GB of memory, a fast SSD for the OS, applications, and project files, and a hard drive for storage of finished projects.

CPU AMD Threadripper 3960X 24-core
Motherboard Gigabyte TRX40 AORUS PRO WIFI
Video Cards 3x NVIDIA GeForce RTX 2060 SUPER
RAM 64GB DDR4
Drives 1TB NVMe SSD + 4TB SATA HDD

Better PC Configuration for V-Ray Next GPU

Our most popular workstations for V-Ray Next GPU bump the video card count up to four and move to using the top model in NVIDIA's GeForce RTX series: the 2080 Ti, with 11GB of graphics memory. That also requires moving to a motherboard with support for more PCI-Express slots and a bigger power supply, and with all of these expensive components a longer warranty is a good idea as well.

CPU Intel Xeon W-2235 8-core
Motherboard Asus WS C422 SAGE/10GB
Video Cards 4x NVIDIA GeForce RTX 2080 Ti
RAM 64GB DDR4 ECC Registered
Drives 1TB NVMe SSD + 4TB SATA HDD

Best PC Configuration for V-Ray Next GPU

Maxing-out GPU rendering performance in V-Ray Next means going up to four of the biggest video cards NVIDIA makes right now, which are the Quadro RTX 8000 that have a massive 48GB of memory each. This may be overkill for most folks, and the RTX 6000 is just as fast but with less VRAM (24GB per card) if you want to save some money.

CPU Intel Xeon W-2295 18-core
Motherboard Asus WS C422 SAGE/10GB
Video Cards 4x NVIDIA Quadro RTX 8000
RAM 256GB DDR4 ECC Registered
Drives 1TB NVMe SSD + 4TB SATA HDD

That is the most powerful traditional, tower workstation that we can build for V-Ray Next GPU right now, but there are other ways to move beyond even this powerhouse of a PC. For example, if this were going to be a server-style system that was accessed remotely, you could go with a rackmount chassis and potentially fit more than four video cards. V-Ray also supports distributed network rendering, so you could split out processing across several systems in either an office or server farm.

Additional Resources

If you want to know more about V-Ray Next performance, we have published several articles over the years looking at how individual GPUs compare as well as how well this software scales across multiple cards. And if you aren’t sure what you need, or if your workflow includes multiple applications, please feel free to call or email our consultants to get a more personalized configuration.

Looking for a V-Ray Workstation?

Puget Systems offers a range of poweful and reliable systems that are tailor-made for your unique workflow.

Configure a System!

Labs Consultation Service

Our Labs team is available to provide in-depth hardware recommendations based on your workflow.

Find Out More!
Tags: Render, Rendering, Workstation, PC Workstation, PC, V-Ray, Chaos, Group, GPU Acceleration, GPU
任柔

If I have a model that needs a particularly large Vram (for example 11G Vram)。If I need to render 100 frames, If ①2*2060 Super Fast or ② 1*2080Ti Fast?
and
if the model need 48G Vram, If ① 2*RTX 6000 24G NVlink Fast or ② 1*RTX 8000 48G Fast ?
my english is bad, hope you know what i mean (╯︵╰)
Sincerely Yours!

Posted on 2020-06-22 15:02:39

Not all applications can utilize NVLink, but V-Ray Next is one of them. I would expect that dual 2060 SUPER cards in NVLink should outperform a single 2080 Ti, and likewise that dual RTX 6000 in NVLink should outdo a single RTX 8000... but please note some caveats:

- While V-Ray Next can use NVLink to allow pairs of video cards to access each other's VRAM, I suspect there is some overhead - so I doubt that you would get a full doubling of effective memory space when using it. I don't have data to back this up, but it is based on my understanding of how NVLink and GPU VRAM function. So a scene that needed 11GB to render would probably be fine on dual RTX 2060 SUPER cards (8GB each) - but I would *not* expect a scene that needed a full 48GB of memory to work on a pair of RTX 6000s; I think that is just cutting it too close. But a scene that needed 30 or 40GB? Yeah, that would likely work :)

- I know you only asked about single vs dual video cards, but for maximum performance some folks want to crank that up to four video cards. NVLink used to be something you could set up to work in multiple pairs within a single system, so if you had four GPUs you could have them in two NVLink pairs. However, sometime in the last year or so NVIDIA seems to have screwed that up... on purpose or by accident, I am not sure. In our latest testing, recent NVIDIA drivers would only allow one pair of cards at a time to be in SLI (which is how to enable NVLink in Windows). You may still be able to work around this, though, using older drivers: https://www.pugetsystems.co...

Posted on 2020-06-22 18:35:56
任柔

Thanks!and i have a new question: What happens if the VRAM is not enough? v-Ray will say“sorry I don't have enough Vram” or Vray still can works but very slow?
(ಥ_ಥ)
eg:model need 20G Vram
① 2*2060 Super NVlink have 16G Vram (2*261=522 score)
②1*Titan RTX have 24G Vram (365score)

Posted on 2020-06-25 06:36:19