Octane for Houdini on Multiple GPUs: Setup, Memory and Versions
Octane for Houdini renders on every GPU you enable, but memory does not add up across cards. What to set before a long render:
- GPUs: Octane Plugin Options › GPUs tab: enable each card
- Speed: Otoy: more GPUs scale render performance roughly linearly
- Memory: VRAM is not combined; out-of-core spills to host RAM at a speed cost
- Version: Octane 2026.5 supports Houdini 22.0.368 and 22.0.429
- License: Studio+: no GPU limit per machine, render node licenses included
Octane is a common choice for Houdini artists who come from Cinema 4D or want a fast unbiased GPU renderer next to Karma. On a multi-GPU machine it can use every card, but only if the plugin is set to, the plugin build matches your Houdini build, and the scene fits on each card or out-of-core is configured.
This guide covers GPU selection in Octane for Houdini, how memory behaves across several GPUs, the out-of-core settings, version matching with Houdini 22, and licensing on a render machine.
Checked against the Otoy Octane for Houdini documentation, Otoy release posts for Octane 2026.5 for Houdini and the Otoy purchase page, November 2026.
Enabling every GPU in Octane for Houdini
Octane for Houdini keeps its settings in the Octane Plugin Options window, opened from the Octane shelf. Its GPUs tab lets you select the GPUs used for rendering when more than one is installed, set render priority to keep a display GPU responsive, and choose per device whether it is used for tone mapping and denoising. Houdini stores these options in the user options folder, so they apply to every Octane scene on that machine.
On a render machine this has one consequence: the GPU selection belongs to the machine, not the .hip file. After installing Octane on a new machine, open Plugin Options and confirm all cards are ticked before the first render.
Sources: Octane for Houdini: Plugin Options, Otoy: Devices tab (Octane for Cinema 4D).
How Octane memory works across several GPUs
Otoy’s device documentation states that adding GPUs to the same machine increases render performance roughly linearly, but VRAM is not compounded the same way: each card needs the scene. On a 4x RTX 5090 server that means 32 GB per card for the in-core scene, not 128 GB.
When a scene does not fit, Octane’s out-of-core mode lets textures and geometry use host RAM. In the Out-Of-Core tab you enable it and set the CPU RAM Usage Limit and the GPU Head Room.
Otoy notes two limits: rendering slows down when textures are streamed from host memory, and out-of-core memory is shared between GPUs, so devices cannot be turned on or off while it is in use. Use out-of-core to make a heavy frame render at all, then reduce texture sizes if the slowdown is large. For volume-heavy scenes see our Octane VDB guide.
Sources: Octane for Houdini: Plugin Options, Otoy: Devices tab.
Matching Octane and Houdini versions
Octane for Houdini is built for specific Houdini builds. Otoy’s release post for OctaneRender 2026.5.0.0 for Houdini lists the supported production builds, including Houdini 22:
Houdini builds supported by Octane 2026.5
From Otoy’s release post for OctaneRender 2026.5.0.0 for Houdini.
| Houdini | Builds | Note |
|---|---|---|
| 22.0 | 22.0.368, 22.0.429 | Windows x64 and Linux (gcc 11.2 builds) |
| 21.0 | 21.0.700, 21.0.729 | Check the post for the exact list per OS |
| 20.5 | 20.5.684 | Older line still supported |
| Install | Houdini package file | Otoy notes houdini.env installs are no longer supported for the Solaris plugin |
| Solaris | Separate OctaneSolaris plugin | Hydra delegate for LOPs; check its own release post |
Since the 2025.1 beta, Octane for Houdini installs through a Houdini package file. Use the build listed for your exact Houdini version on every machine that renders, or the scene may not load the plugin. If you also render Karma or Redshift in the same project, our Houdini 22 on a render farm guide covers version matching for all of them.
Sources: Otoy: OctaneRender 2026.5.0.0 for Houdini, Otoy: OctaneRender for HoudiniSolaris 2026.2.
Licensing and network rendering
Octane is licensed by Otoy, separately from Houdini. Otoy’s purchase page says all Studio+ subscriptions include ten or more Render Node licenses with no GPU limits per machine. Older Studio licenses were limited to two GPUs, so if a third GPU is refused on a machine, check which license and which build is installed.
For rendering one frame across several machines, the Octane ROP has Network Options to make the machine a primary render node, and Otoy’s docs describe automatic port configuration so several primary nodes can run on one computer. On a single multi-GPU server you do not need network rendering: enable all GPUs locally.
Sources: Otoy: OctaneRender purchase, Octane for Houdini: Enabling Network Rendering.
SaaS or IaaS render farm for Octane in Houdini?
Which farm type fits Octane for Houdini
Octane needs your Otoy license; the farm type decides how you bring it.
| Situation | Usually a better fit | Why |
|---|---|---|
| Farm supports your exact Octane and Houdini builds | SaaS (managed farm) | Submit if the farm’s plugin build matches yours |
| You need a specific Octane build for your Houdini build | IaaS (remote workstation) | Install exactly the build you tested |
| Scene needs out-of-core tuning | IaaS | Adjust RAM limit and head room on the machine that renders |
| Octane and Karma in one project | IaaS | Both renderers installed side by side, same project paths |
| Short one-off job | SaaS | No setup time if versions match |
Rule of thumb: if a managed farm runs your exact Octane and Houdini builds, it is the least work. Otherwise a machine you control lets you install the matching builds and sign in with your own Otoy license.
Rendering Octane for Houdini on an iRender server
On iRender you rent a whole Windows workstation, install Houdini and the Octane plugin build that matches it, and sign in with your own SideFX and Otoy licenses:
- Upload the project, textures and caches through iRender Drive with relative paths.
- Start a multi-GPU server; Octane uses every GPU you enable in Plugin Options.
- Install your Houdini build and the matching Octane for Houdini package.
- Open Octane Plugin Options, tick every GPU, and set out-of-core only if the scene needs it.
- Render one frame from the Octane ROP, check the log, then render the range.
Octane for Houdini checklist
Use this as a final check before you start the full range on any machine.
Frequently asked questions
1. How do I make Octane use all GPUs in Houdini?
Open Octane Plugin Options from the Octane shelf, go to the GPUs tab and enable every GPU. The setting is stored per machine.
2. Does Octane combine VRAM across GPUs?
No. Otoy states performance scales with more GPUs but VRAM is not compounded; each card needs the scene unless out-of-core is used.
3. Does Octane support Houdini 22?
Yes. OctaneRender 2026.5.0.0 for Houdini supports Houdini 22.0.368 and 22.0.429, as well as 20.5 and 21.0 builds.
4. What does Octane out-of-core do?
It lets textures and geometry use host RAM when VRAM is full. Rendering slows while textures stream from host memory.
5. How many GPUs can an Octane license use?
Otoy’s purchase page says Studio+ includes 10+ Render Node licenses with no GPU limits per machine. Older Studio licenses were limited to two GPUs.
6. Where are Octane GPU settings stored?
In the Houdini user options folder, so they apply to all Octane scenes on that machine.
7. Can Octane render one frame across several machines from Houdini?
Yes, with network rendering: the Octane ROP’s Network Options make a machine a primary render node that uses render nodes.
Render your Houdini scenes on iRender
iRender is an IaaS GPU render farm: you rent a bare-metal server with 4x NVIDIA RTX 5090 (32 GB) 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. You install Houdini and use your own SideFX license; your server image is saved between sessions, so Houdini, your packages and the Karma XPU kernel cache stay installed. For Karma XPU we recommend the 4-GPU servers.
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 servers are liquid-cooled; in our own measurements the GPUs hold 60–70°C under sustained render load. Elite support is available 7 days a week by hotline, live chat, Skype, WhatsApp and email.
This video walks through rendering a Houdini Karma XPU scene on an iRender server in five steps:
New users get a 100% bonus on the first top-up (minimum $50). See the Karma XPU render farm page for server options, or download the iRender GPU app.
If Karma XPU renders are slow on your machine, iRender’s Karma XPU render farm gives you a dedicated remote server where you install Houdini, bring your own license and render on multiple RTX 5090 GPUs (for Karma XPU, 4 GPUs is usually the most efficient setup).


