Hey,
We upgraded to the latest C4D and Redshift versions and we encounter the problem that running multiple workers, Redshift oder C4D are pushing the VRAM to 99% utilization and there for the workers are all running out of VRAM, resulting in endless errors.
Does anyone also has this problems with the latest version. Its mostly caused by Redshift because 2026.4 worked but 2026.5 doesnt. We also tried CPU and GPU affinity settings, no change in behaviour. Redshift supports says its not their issue because commandline renderer itself works, which I cant confirm tho
can you compare running redshift outside of deadline and using a single card, is this the same command in deadline?
-device followed by the device ordinal N enables that device for rendering
For example, to render scene 'test.rs' with the first two devices, do:
redshiftCmdLine test.rs -device 0 -device 1
Use -device all to enable all devices
If -device is not specified, the devices specified in the preferences.xml are used
-gpu is an alias of -device used for backward compatibility
Has this in the docs, not sure if anything changed or this needs updating in the deadline files
So I also checked this with the Redshift Forum: When I run a single instance with the commandline renderer, it works, that also applies to Deadline.
The problem occures with multiple instances. In Deadline as well as Commandline Renderer directly. But its all with 2026.5 Redshift. 2026.4 does all as expected.
Here is also the error log that I get from Deadline, Commandline Renderer always stalls and doesnt produce logs. The Deadline Log looks the same across all workers
2026-05-05 15:59:26: 0: STDOUT: Redshift Debug: 0ms
2026-05-05 15:59:26: 0: STDOUT: Redshift Debug:
2026-05-05 15:59:26: 0: STDOUT: Redshift Debug: Loading OCIO config using C:\Program Files\Maxon Cinema 4D 2026\resource\modules\c4d_base\ocio\config.ocio
2026-05-05 15:59:26: 0: STDOUT: Redshift Debug: Full path: C:\Program Files\Maxon Cinema 4D 2026\resource\modules\c4d_base\ocio\config.ocio
2026-05-05 15:59:26: 0: STDOUT: Redshift Debug: Ok
2026-05-05 15:59:26: 0: STDOUT: Redshift Debug: Creating OCIO processors between rendering and view/srgb spaces
2026-05-05 15:59:26: 0: STDOUT: Redshift Debug: Rendering space: scene-linear Rec.709-sRGB
2026-05-05 15:59:26: 0: STDOUT: Redshift Debug: Display: sRGB
2026-05-05 15:59:26: 0: STDOUT: Redshift Debug: View: Un-tone-mapped
2026-05-05 15:59:26: 0: STDOUT: Redshift Debug: Ok
2026-05-05 15:59:26: 0: STDOUT: Redshift Debug: Loading OCIO color space transforms for texture sampling
2026-05-05 15:59:26: 0: STDOUT: Redshift Debug: Found a suitable sRGB color space: "sRGB"
2026-05-05 15:59:26: 0: STDOUT: Redshift Debug: Found a suitable sRGB-linear color space: "scene-linear Rec.709-sRGB"
2026-05-05 15:59:26: 0: STDOUT: Redshift Info:
2026-05-05 15:59:26: 0: STDOUT: Redshift Info: Device 0 (NVIDIA GeForce RTX 4090) uses OptiX RT for ray tracing
2026-05-05 15:59:26: 0: STDOUT: Redshift Info: Device 1 (NVIDIA GeForce RTX 4090) uses OptiX RT for ray tracing
2026-05-05 15:59:26: 0: STDOUT: Redshift Info: Device 2 (NVIDIA GeForce RTX 4090) uses OptiX RT for ray tracing
2026-05-05 15:59:26: 0: STDOUT: Redshift Info: Device 3 (NVIDIA GeForce RTX 4090) uses OptiX RT for ray tracing
2026-05-05 15:59:26: 0: STDOUT: Redshift Info: Device 4 (NVIDIA GeForce RTX 4090) uses OptiX RT for ray tracing
2026-05-05 15:59:26: 0: STDOUT: Redshift Info: Device 5 (NVIDIA GeForce RTX 4090) uses OptiX RT for ray tracing
2026-05-05 15:59:26: 0: STDOUT: Redshift Debug: Detected 6 new texture color spaces in use.
2026-05-05 15:59:26: 0: STDOUT: Redshift Debug:
2026-05-05 15:59:26: 0: STDOUT: Redshift Debug: Preparing ray tracing hierarchy for meshes
2026-05-05 15:59:26: 0: STDOUT: Redshift Debug: Time to process 29 meshes: 2ms
2026-05-05 15:59:26: 0: STDOUT: Redshift Debug: Using streamed texture page cache.
2026-05-05 15:59:26: 0: STDOUT: Redshift Debug:
2026-05-05 15:59:26: 0: STDOUT: Redshift Debug: Processing 11 textures
2026-05-05 15:59:26: 0: STDOUT: Redshift Debug: Time to process textures: 0.007384 seconds
2026-05-05 15:59:26: 0: STDOUT: Redshift Debug:
2026-05-05 15:59:26: 0: STDOUT: Redshift Debug: Preparing materials and shaders
2026-05-05 15:59:40: 0: STDOUT: Redshift Debug: Time to process all materials and shaders: 13.338744 seconds
2026-05-05 15:59:41: 0: STDOUT: Redshift Debug:
2026-05-05 15:59:41: 0: STDOUT: Redshift Debug: Pre-processing for Tesselation Free Displacement
2026-05-05 15:59:41: 0: STDOUT: Redshift Debug: Time: 0ms
2026-05-05 15:59:41: 0: STDOUT: Redshift Debug:
2026-05-05 15:59:41: 0: STDOUT: Redshift Debug: Freeing VRAM because of an OptiX denoiser/RT reset
2026-05-05 15:59:41: 0: STDOUT: Redshift Debug: Freeing RT related buffers
2026-05-05 15:59:41: 0: STDOUT: Redshift Debug:
2026-05-05 15:59:41: 0: STDOUT: Redshift Debug: Initializing OptiX RT,
2026-05-05 15:59:41: 0: STDOUT: Redshift Debug: Freeing stack allocator
2026-05-05 15:59:44: 0: STDOUT: Redshift Debug: [Device 0, 1, 2, 3, 4, 5] will rebuild everything
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Succeeded to init OptiX RT on device 0 but too little VRAM left. Will have to disable OptiX RT!
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Freeing OptiX RT, scene
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Time: 7841ms
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug:
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Allocating GPU mem...(device 0)
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Done (Allocator size: 304 MB. CUDA reported free mem before: 816 MB, after: 816 MB)
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug:
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Allocating GPU mem...(device 1)
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Done (Allocator size: 540 MB. CUDA reported free mem before: 1052 MB, after: 1052 MB)
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug:
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Allocating GPU mem...(device 2)
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Done (Allocator size: 728 MB. CUDA reported free mem before: 1240 MB, after: 1240 MB)
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug:
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Allocating GPU mem...(device 3)
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Done (Allocator size: 538 MB. CUDA reported free mem before: 1050 MB, after: 1242 MB)
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug:
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Allocating GPU mem...(device 4)
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Done (Allocator size: 188 MB. CUDA reported free mem before: 700 MB, after: 700 MB)
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug:
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Allocating GPU mem...(device 5)
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Done (Allocator size: 96 MB. CUDA reported free mem before: 608 MB, after: 608 MB)
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug:
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Ray Tracing Hierarchy Info:
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Max depth: 128. MaxNumLeafPrimitives: 8
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Extents: (-24.274837 -32.232719 -8.959898) - (97.304062 133.773727 80.913330)
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Time to create tree: 0 ms (0 0 0)
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug:
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Preparing Tesselation Free Displacement for rendering
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Time: 18ms
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug:
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Irradiance point cloud...
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: [4] Allocating memory for device 0 (NVIDIA GeForce RTX 4090)
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Redshift can use up to 304 MB
2026-05-05 15:59:49: 0: STDOUT: Redshift Error: There's less than 128MB of free VRAM once fixed data (photon map, irradiance point cloud, irradiance cache) are considered. Aborting the render
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Information for out-of-memory situation
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: =======================================
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Stage: IPC
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: ActiveDeviceID: 0 (of 6)
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: System device ordinal: 0 (NVIDIA GeForce RTX 4090)
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: FreeMemOriginally: 304 MB
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Fixed data (includes volume grids, instance matrices, textures): 544 MB
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Photons: 0 bytes
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Irradiance Point Cloud: 0 bytes
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Irradiance cache: 0 bytes
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: SSS: 0 bytes
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Textures: 3612 bytes
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug: Rays: 0 MB
2026-05-05 15:59:49: 0: STDOUT: Redshift Debug:
2026-05-05 15:59:53: 0: STDOUT: Redshift Debug: Total irradiance point cloud construction time 4.2s
2026-05-05 15:59:53: 0: STDOUT: Redshift Info:
2026-05-05 15:59:53: 0: STDOUT: Redshift Info: Rendering time: 27.5s (6 device(s) used)
This looks like it’s using all 6 cards, if you have several workers you’ll need to set the affinity on each worker. I wouldn’t recommend using 6 workers, probably 3 workers with 2 cards each (or 2 with 3).
If you’re running multiple instances on the same worker this will probably fail as it won’t separate the cards.
You can test outside of deadline by running the benchmark multiple times in separate command prompt windows usine -gpu/device options
My setup previously was 3x2 GPUs, just wanted to test out a 6x1 Setup,
But I am certain that I set up everything correctly because again 2026.4 works flawless and previous versions also did.
Its more like Maxon/Redshift Team isnt really a help currently because they are not investigating this at the moment. So I was definitly wondering if more people have to issue
2026.6 Update didnt change anything, it ignores the GPU affinity options that I set for every worker I have
i didn’t see anything in the release notes either.
Did it work outside of Deadline? I don’t have a multicard machine to test this on, but I’d run a test to see if this is working outside of deadline(which version are you on)?
I got confirmation from Maxon that they investigate this issue right now. Suspected Commandline Renderer Bug
1 Like
We have confirmed that this issue is related to Cinema 4D and Command Line version 2026.2.0.
While this is being addressed, there are two available workarounds:
Roll back to Cinema 4D 2026.1.4 and Redshift 2026.4.
Roll back to Cinema 4D 2025.3.3, which allows you to continue using the latest version of Redshift.
For context on why the second option works, the Redshift 2026.5 release notes specify that the minimum requirement for the 2026 plugin has increased to Cinema 4D 2026.2. Because this restriction only applies to the 2026 plugin, Cinema 4D 2025 remains fully compatible with the latest Redshift releases.
Official Maxon Support Statement
2 Likes
Did they fix the issue with this release?
Version 2026.6.2 is now available [2026.05] | Redshift Forums
RELEASE NOTES
2026.6.2 (2026.05)
- [C4D] Fixed a bug where the device selection via the CommandLine did not work