Octane Multi-GPU Rendering in Cinema 4D: Devices, Priority and VRAM
Octane renders one frame on every enabled GPU at once. Check these before a multi-GPU render:
- Devices tab: tick each GPU that should render and set it to High priority
- Final render: the Picture Viewer uses the same GPUs unless Use All GPUs is on
- Memory: VRAM is not combined; the smallest active card sets the in-core limit
- Out-of-core: textures and geometry can move to system RAM, shared by all GPUs, at a speed cost
- License: OTOY’s Studio+ plan has no limit on the number of GPUs
Octane is one of the GPU renderers that gets faster with almost every card you add. OTOY’s own documentation says performance grows roughly linearly with more GPUs, which is why Cinema 4D artists move Octane jobs to 4-GPU and 8-GPU machines. The catch is that a few settings decide whether Octane actually uses those cards, and memory works very differently from speed.
This guide explains how Octane splits a frame across GPUs in Cinema 4D, where you choose the devices, why a multi-GPU machine can still render on one card, and how to plan memory when the cards are not identical.
Checked against the OctaneRender for Cinema 4D manual (Devices tab, Main settings, Out of Core tab) and OTOY Help Center articles, November 2026.
1. How does Octane split one frame across several GPUs?
Octane does not give each GPU its own frame. It sends the scene to every enabled GPU, and each card renders part of the samples for the same frame; the samples are added together into one image. That is why one frame of a still image or an animation gets faster as you add GPUs.
OTOY’s Devices tab page describes the scaling as roughly linear, and adds the important limit in the same place: every GPU needs a full copy of the scene assets it processes. Speed adds up; memory does not. A machine with four 32 GB cards still renders scenes that fit in 32 GB, only faster.
2. Where do you choose the GPUs for Octane in Cinema 4D?
Open the Live Viewer and go to Settings > Devices. The tab lists every GPU in the machine with check boxes for rendering, render priority, imaging (tone mapping) and denoising. If a box is disabled, that GPU will not be used for that function. The Device Settings button opens the OctaneRender device preferences, which repeat some of these settings and add Use RTX acceleration, enabled by default.
Render Priority works like CPU priority in Task Manager. OTOY’s advice depends on what drives your monitor: on a workstation, give the card that runs the display a lower priority, or leave it out of rendering on heavy scenes, and set the other cards to High. On a machine with several identical GPUs, OTOY recommends enabling all of them at High priority and not mixing in a weaker card.
Live Viewer devices vs the final render
The Devices tab belongs to the Live Viewer, but the Octane render settings in Cinema 4D have one more switch on the Main tab: Use All GPUs. When it is on, a final render uses every GPU in the machine, including cards you turned off in the Live Viewer. It is off by default, so a card you unticked to keep the interface smooth will also sit idle in the Picture Viewer and Render Queue.
On a render machine where nobody is working in the interface, the simplest setup is to tick every GPU in the Devices tab at High priority. On your own workstation, keep one card for the display while you work, then turn on Use All GPUs for the overnight render.
3. Why do four GPUs not give you four times the VRAM?
Because each GPU renders the whole frame’s samples, each one needs the scene data it touches. OTOY states that the usable memory is limited to the card with the least VRAM in the machine or network. With identical cards that is simply the memory of one card. With mixed cards, the smaller one decides.
OTOY’s advice for large scenes is to disable GPUs that do not have enough VRAM, so the scene can fit in the remaining cards. Two more things change the budget. RTX acceleration, on by default, uses additional VRAM for geometry, as OTOY support notes on its forum. And out-of-core lets Octane place textures and geometry in system RAM when they no longer fit: Enable Out of Core Textures/Geometry and a RAM Usage Limit sit in the Out of Core tab. OTOY points out that out-of-core memory is shared between GPUs, so you cannot switch devices on or off while it is in use, and RAM you give it is no longer available to other programs.
For the full memory picture, see Octane out-of-core in Cinema 4D and mixing an RTX 5090 with an RTX 4090.
4. Why is Octane only using one GPU?
When a multi-GPU machine renders no faster than a single card, the cause is usually one of a handful of settings. The table lists them in the order worth checking.
Why Octane renders on fewer GPUs than you have
Work down the rows; most cases are a setting, not a hardware fault.
| Symptom | Likely cause | What to do |
|---|---|---|
| Only one GPU busy in the Live Viewer | Other GPUs unticked in the Devices tab | Tick every GPU you want to render with and set it to High priority |
| Live Viewer uses all GPUs, final render does not | Picture Viewer follows the Devices tab selection | Tick the GPUs again or turn on Use All GPUs in the Octane render settings |
| GPU listed but greyed out for a function | That GPU is disabled for rendering, imaging or denoising | Enable the box for the function you need |
| GPUs look idle in Task Manager | The default graphs show 3D and video engines, not compute | Check per-GPU use and memory with nvidia-smi |
| Busy only for a moment each frame | The CPU is preparing and sending the scene | Look at scene export time; see the slow-frames guide |
| A card drops out on heavy scenes | Scene does not fit in that card’s memory | Turn off the smaller card or reduce memory use |
Two of these trip people up most. The first is the Picture Viewer following the Live Viewer selection, covered above. The second is reading GPU use from Windows Task Manager: its default GPU graphs track 3D and video engines, so a card rendering CUDA work can look idle. NVIDIA’s nvidia-smi tool reports utilisation and memory per GPU and is the more reliable check.
If every card shows activity only in short bursts, the GPUs are waiting for Cinema 4D to prepare the next frame. That is a CPU-side export issue rather than a device setting.
Watching every GPU during a render
The quickest proof that every card works is to watch them during a real render. The first command prints one line per GPU every two seconds. While a frame renders, each enabled GPU should show high utilisation and power draw, and similar memory use, because each holds its own copy of the scene. A card that stays near 0% while the others work is not enabled for that render; a card whose memory reaches its total is the one that will fail first on a heavier frame.
For background renders, OTOY documents an -octaneGPU argument for the Commandline application that switches each device on or off by number, for example "0:true 1:false". The second command uses it to render on four GPUs.
OTOY notes that background rendering with Commandline needs a valid command line license. The numbers may not line up between tools: nvidia-smi orders cards by PCI bus, while CUDA applications by default number the fastest card first (NVIDIA’s CUDA_DEVICE_ORDER). On a server with identical cards, match them by the names and memory shown in Octane’s Devices tab and log.
Sources: Dell: useful nvidia-smi queries, OTOY: Command Line Rendering, NVIDIA: CUDA environment variables.
5. Licensing Octane for multi-GPU machines
Octane is licensed by OTOY, separately from Cinema 4D. OTOY’s subscription page states that Studio+ has no limit on the number of GPUs you can use, and it includes ten render node licenses for Octane’s own network rendering, with more available on request. Render managers such as Deadline or Team Render are different: OTOY says each render node there needs its own Studio+ license.
For Cinema 4D on Windows, OTOY’s current plugin downloads are labelled Studio+ and Demo; the free Prime build of the C4D plugin is for macOS. Use a Studio+ account for multi-GPU Windows machines.
6. SaaS or IaaS render farm for multi-GPU Octane jobs?
Which farm type fits a multi-GPU Octane render
Both can render Octane on several GPUs; they differ in who controls the device settings.
| Situation | Usually a better fit | Why |
|---|---|---|
| Scene already renders cleanly on all your GPUs | SaaS (managed farm) | Submit the file; the farm handles devices |
| Device setup or memory still uncertain | IaaS (remote workstation) | Open the Live Viewer and check every card yourself |
| Scene close to the VRAM of one card | IaaS with 32 GB cards | Watch per-GPU memory on the heaviest frame |
| Need a specific Octane plugin build | IaaS | Install the exact build you use at home |
| Short one-off render | SaaS | No setup time |
Rule of thumb: if your scene already renders correctly on several GPUs at home, either type works. If you still need to check devices, priority or memory on the target machine, a remote workstation you control saves failed submissions.
Rendering Octane on an iRender multi-GPU server
On iRender you rent a whole Windows workstation with Cinema 4D and OctaneRender installed, and sign in with your own OTOY account:
- Upload the project with all textures and caches through iRender Drive before you start the server.
- Start a 2x, 4x or 8x RTX 5090 server (32 GB per card). All cards in a server are the same model, so the in-core limit is 32 GB.
- Open the Live Viewer, go to Settings > Devices and check that every GPU is ticked at High priority.
- Render the heaviest frame and watch memory per GPU with nvidia-smi.
- Start the full range in the Picture Viewer or Render Queue; turn on Use All GPUs if any card is unticked.
Octane multi-GPU checklist
Use this as a final check before you start the full range on any machine.
Frequently asked questions
1. Does Octane use multiple GPUs in Cinema 4D?
Yes. Octane renders each frame on every enabled GPU at the same time, with each card computing part of the samples. OTOY describes the speed gain as roughly linear with more GPUs.
2. Does Octane combine the VRAM of several GPUs?
No. Each GPU needs a copy of the scene data it renders, so the usable in-core memory is set by the card with the least VRAM. Out-of-core can move textures and geometry to system RAM instead, at a speed cost.
3. Why does the Picture Viewer use fewer GPUs than the Live Viewer?
A final render uses the GPUs ticked in the Live Viewer Devices tab. If you unticked a card, it stays idle unless you turn on Use All GPUs in the Main tab of the Octane render settings.
4. What does Render Priority do in Octane?
It controls how hard Octane uses each GPU, similar to CPU priority. Lower it on the card that drives your monitor to keep the interface responsive, and use High on cards that only render.
5. Should I mix a weaker GPU with stronger ones for Octane?
Only if the scene fits in the weaker card’s memory. OTOY recommends disabling GPUs that do not have enough VRAM for large scenes, and not adding a weaker card to a set of identical GPUs.
6. Does RTX acceleration change VRAM use?
Yes. OTOY support notes that RTX acceleration increases the VRAM used for geometry. It is enabled by default in the device settings and usually speeds up rendering, so leave it on unless memory is the limit.
7. How many GPUs can an Octane license use?
OTOY states that Studio+ has no limit on the number of GPUs. It also includes ten render node licenses for Octane network rendering; Deadline or Team Render nodes each need their own Studio+ license.
8. How can I see whether every GPU is rendering?
Use nvidia-smi, which reports utilisation and memory per GPU. The default Windows Task Manager graphs can show a compute-only card as idle.
Render your Cinema 4D + Octane scenes on multi-GPU servers
iRender is an IaaS GPU render farm: you rent a bare-metal server with 1, 2, 4 or 8 NVIDIA RTX 4090 or RTX 5090 GPUs, an AMD Ryzen™ Threadripper™ PRO CPU, 256 GB ECC RAM and a 2 TB NVMe SSD, connect by remote desktop and work on it like your own workstation, with full administrator access. Cinema 4D and OctaneRender are installed; sign in with your own OTOY account (OTOY’s Studio+ plan has no GPU limit for multi-GPU servers). Your server image is saved between sessions.
Multi-GPU servers are billed per minute (1-GPU servers have a one-hour minimum), and sessions of 3 hours or more get 10% off. Our 4-GPU and 8-GPU servers are liquid-cooled; in our own measurements the GPUs hold 60–70°C under sustained render load, so they keep their boost clocks. Elite support is available 7 days a week by hotline, live chat, Skype, WhatsApp and email.
This quick start shows how to start Cinema 4D and Octane on an iRender server:
On a 4x RTX 5090 server, a 300-frame Cinema 4D + Octane animation rendered in 1 hour 18 minutes. The second clip runs the same kind of scene on 4x RTX 5090 and 4x RTX 4090:
New users get a 100% bonus on the first top-up (minimum $50). See plans on the Octane render farm page, download the iRender GPU app, or read mixing an RTX 5090 with an RTX 4090.
Sources: OTOY: Devices tab (C4D), OTOY: Main settings, Use All GPUs (C4D), OTOY: Out of Core tab (C4D), OTOY: Hardware options, disabling small GPUs, OTOY forum: RTX acceleration and VRAM, OTOY: OctaneRender Subscriptions, OTOY: Octane 2026.4 for Cinema 4D release, NVIDIA: nvidia-smi. Checked November 2026.




