Octane Out-of-Core in Cinema 4D (2026): Textures, Geometry, RAM Limits and Why a Scene Still Runs Out of Memory
When a scene does not fit in GPU memory, Octane can keep part of it in system RAM. OTOY's documentation and its memory guides show what out-of-core covers, what it costs and where it stops.
- Textures and geometry. Current Octane for Cinema 4D has Enable Out of Core Textures/Geometry; out-of-core geometry arrived with Octane 4.
- Slower, not free. OTOY: rendering with out-of-core data is slower than with everything in VRAM.
- Locked RAM. Out-of-core data sits in non-swappable memory that other applications cannot use.
- Still has a ceiling. If the out-of-core budget is exceeded, Octane reports it is out of memory or the render fails.
- Measure first. Device Settings shows how much VRAM the scene and engine actually use.
“Out of memory” is the most common way a heavy Octane scene stops. Octane’s out-of-core feature lets it borrow system RAM when the GPU is full, but it has settings, costs and limits that decide whether the render finishes. This guide follows OTOY’s Octane for Cinema 4D documentation and OTOY’s memory guides.
Checked against the OTOY documentation, November 2026.
1. What does Octane out-of-core actually do?
OTOY describes out-of-core as letting OctaneRender use the computer’s main system memory when scene assets no longer fit in the available VRAM. In current Octane for Cinema 4D the switch is Enable Out of Core Textures/Geometry, and OTOY notes that with out-of-core geometry support a significant portion of system memory can be used for geometry, also with multiple GPUs. OTOY’s Standalone documentation dates out-of-core geometry to Octane 4.
Some older OTOY pages for other hosts still say out-of-core is for textures only; those pages describe earlier versions.
2. How to set up out-of-core in Cinema 4D (step by step)
- Open the Out of Core tab in the Octane settings and enable Out of Core Textures/Geometry.
- Set the RAM Usage Limit (GB). This is how much system memory Octane may reserve. Leave enough for Cinema 4D and the operating system, because OTOY notes this memory is non-swappable and unavailable to other applications.
- Keep GPU Head Room near its default. It is the GPU memory left free on each card; OTOY’s Standalone notes keep it small because VRAM is faster than system RAM.
- Set devices first. OTOY notes out-of-core memory is shared between GPUs, so you cannot turn devices on or off while using it.
- Render and check Device Settings (Octane Settings › Settings tab › Device Settings) to see how much VRAM and out-of-core memory the scene uses.
| Setting | What it does (OTOY) | Tip |
|---|---|---|
| Enable Out of Core Textures/Geometry | Lets Octane keep textures and geometry in system RAM when VRAM is full | Turn on only when the scene does not fit |
| RAM Usage Limit (GB) | Caps how much system memory out-of-core may use | Leave room for C4D and the OS |
| GPU Head Room | GPU memory kept free on each card | Keep near default |
| Device Settings | Shows VRAM use per category after a render | Check before and after changes |
Source: OTOY, Out of Core tab (Cinema 4D) and Out of Core settings (Standalone).
On network render nodes, OTOY notes each node also needs enough RAM, and the out-of-core amount is set when installing the node daemon.
3. What OTOY measured
OTOY’s Help Center ran a series of memory tests in Cinema 4D. The figures below are OTOY’s, not iRender measurements, and they depend on the scene and hardware.
| Test in OTOY's guides | Result reported |
|---|---|
| Empty-ish scene, Device Settings | About 1.1 GB of VRAM in "Engine Runtime Data" (overhead) |
| Two million triangles | About 367 MB of VRAM once sent to Octane |
| 64 million quads (about 128 million triangles) | Rendered with 20 GB of out-of-core memory, after 13 minutes of pre-processing |
Source: OTOY Help Center, Overview, Polygons and Findings Summary (Cinema 4D memory series).
Two practical points follow. Cinema 4D’s quads and n-gons are converted to triangles before Octane renders them, so polygon counts in Cinema 4D understate what Octane stores. And very heavy out-of-core geometry can take minutes to prepare before the first sample, per frame, so check it on a test frame before committing an animation.
4. Why does my scene still run out of memory?
The cheapest fix is often in the scene: OTOY’s memory series shows how subdivision levels multiply polygon counts, and render-time subdivision and displacement add geometry Cinema 4D’s own counter does not show.
5. More VRAM or more out-of-core?
Out-of-core keeps a render alive; more VRAM keeps it fast. A card with more memory fits more of the scene in VRAM, so less of it has to come from system RAM. VRAM is not pooled across GPUs: each GPU holds its own copy of the scene, so a multi-GPU machine needs the scene to fit on each card, with out-of-core RAM shared between them.
On iRender servers each RTX 5090 has 32 GB of VRAM and each RTX 4090 has 24 GB, with 256 GB of system RAM, so a large out-of-core budget is available when a scene needs it. We have not timed an out-of-core scene on our servers yet; the table lists Cinema 4D + Octane runs recorded in iRender videos.
Octane runs measured on iRender servers
| Test scene | Server | Frames | Render time | Cost at list price | Video |
|---|---|---|---|---|---|
| Car drift scene1080p, 800 max samples | 5i 4x RTX 5090 | 300 | 1h 18m 24s · 15.6 s/frame | ~$49.65 | Watch the test |
| Car drift scenesame scene and settings | 5S 4x RTX 4090 | 300 | 18.8 s/frame (about 1h 34m) | ~$47 (estimated from s/frame) | Watch the test |
| Cinema 4D animation1080p | 9S 8x RTX 4090 | 320 | 1h 02m render · 1h 16m session billed | ~$65.87 | Watch the test |
| Sailboat animation1080p | 5i 4x RTX 5090 | 120 | 2h 01m | ~$76.63 | Watch the test |
The rows are different scenes and do not rank the servers. Render one frame of your heavy scene and read Device Settings to see whether it fits in VRAM or relies on out-of-core.
A test workflow for heavy scenes on an iRender server
- Upload the project and textures to iRender Drive before you start the server, so upload time is not billed.
- Start a 4x or 8x RTX 5090 server (packages 5i or 9i); multi-GPU servers are billed per minute. Cinema 4D and OctaneRender are pre-installed; sign in with your own OTOY account. OTOY’s Studio+ plan has no GPU limit.
- Set devices, enable out-of-core if needed and set the RAM Usage Limit before rendering.
- Render one frame and check Device Settings and the pre-processing time.
- Render the sequence; split long frame ranges across several servers.
FAQ
1. Does Octane out-of-core work for geometry?
Yes. Current Octane for Cinema 4D has Enable Out of Core Textures/Geometry, and OTOY added out-of-core geometry in Octane 4. Some older OTOY pages for other hosts still say textures only.
2. Why do I still get out of memory with out-of-core on?
The out-of-core budget has a limit set by RAM Usage Limit. If geometry and textures exceed VRAM plus that budget, Octane reports it is out of memory.
3. Does out-of-core slow down Octane?
Yes. OTOY notes rendering with out-of-core data is slower than with everything in VRAM, and very heavy geometry can take minutes to pre-process.
4. How much RAM should I give out-of-core?
Enough for the overflow, but leave memory for Cinema 4D and the operating system, because out-of-core RAM is non-swappable and unavailable to other applications.
5. How do I see how much VRAM my Octane scene uses?
After a render, open Octane Settings › Settings tab › Device Settings for a breakdown of VRAM use.
6. Does VRAM add up across GPUs in Octane?
No. Each GPU holds the scene; out-of-core memory is shared between GPUs, but VRAM is not pooled.
7. Can I render heavy Octane scenes on iRender?
Yes. iRender servers have 1 to 8 RTX 5090 (32 GB) or RTX 4090 (24 GB) GPUs and 256 GB of RAM, with Cinema 4D and OctaneRender pre-installed; sign in with your own OTOY account.
Render your heavy Cinema 4D + Octane scenes on iRender
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, connect by remote desktop and work on it like your own workstation.
- Multi-GPU servers: 1x to 8x RTX 4090 (24 GB) or RTX 5090 (32 GB), AMD Threadripper PRO CPUs, 256 GB ECC RAM and 2 TB NVMe SSD.
- Octane ready: OctaneRender is installed on the servers; you sign in with your own OTOY account and open your project in Cinema 4D as usual. Your server image is saved between sessions, and the physical machine is fully wiped before anyone else uses it.
- Pay for what you use: billed per minute (1-hour minimum on single-GPU servers), with 10% off sessions of 3 hours or more. See pricing.
- Elite Support, 7 days a week: by hotline, live chat, Skype, WhatsApp and email.
The quickstart below shows how to render a Cinema 4D + Octane project on an iRender server, from uploading the files to downloading the frames.
Two iRender tests show what the 4-GPU servers do with a Cinema 4D + Octane animation. In the first, 4x RTX 5090 rendered 300 frames in 1 h 18 min; the second compares 4x RTX 5090 with 4x RTX 4090 on the same scene.
New accounts get a 100% bonus on the first top-up (minimum $50). See the Octane render farm page for server options, or download the iRender GPU app.




