Hi
Can someone confirm what the maximum render resolution is for Octane on a 4GB GTX 980 card?
thx
Maximum render size (print res)
Forum rules
For new users: this forum is moderated. Your first post will appear only after it has been reviewed by a moderator, so it will not show up immediately.
This is necessary to avoid this forum being flooded by spam.
For new users: this forum is moderated. Your first post will appear only after it has been reviewed by a moderator, so it will not show up immediately.
This is necessary to avoid this forum being flooded by spam.
You will probably be limited by the vram of the GPU.
An 8k x 8k render will take up 1.3GB for the render target alone (leaving 2.7GB for the rest of your scene and UI display of the OS).
In theory the max i think is something like 32k x 32k but for this your looking at something like 20GB for the render target so it's not possible at the moment.
What res do you need?
An 8k x 8k render will take up 1.3GB for the render target alone (leaving 2.7GB for the rest of your scene and UI display of the OS).
In theory the max i think is something like 32k x 32k but for this your looking at something like 20GB for the render target so it's not possible at the moment.
What res do you need?
it seems that it's locked at 8k Square, but..You can offset camera & render like 4 tiles (or more) with light overlap & then connect those in post =) this is actually very good technique if You need very high res prints & don't have enough vRAM or face software limitations (just don't use post processing in Octane as You'll have some problems to connect those tiles =)
thanks for the info. I'm a print guy and regularly render at sizes over 20k wide. Minimum is usually approx 10k x 7k.
I use vray, so I'm evaluating whether an offset camera workflow will cause too much of a hassle, I'm guessing it will, as I use a lot of object ID passes as well as beauty pass. I'm cringing just thinking about it.
Anyone know if v3 will support buckets or system ram and bigger buffers?
I use vray, so I'm evaluating whether an offset camera workflow will cause too much of a hassle, I'm guessing it will, as I use a lot of object ID passes as well as beauty pass. I'm cringing just thinking about it.
Anyone know if v3 will support buckets or system ram and bigger buffers?
V3 will be able to use system RAM for textures and geo, but I believe the frame buffer still needs to fit into the cards memory. That shouldn't be a problem. I've rendered a 15K image for print with Octane and using 4GB cards before there was out-of-core support, but it wasn't a complex scene obviously. A 12GB Titan X or 16GB AMD card (once V3 supports OpenCL) should be able to hold a large frame buffer plus a good sized scene before moving assets over to system RAM.
thanks. I have a 12GB Titan X coming next week.
Windows 10 - 64GB RAM - Cinema 4D R20 - RTX 2070 x3
That's not correct. Version 3 will not have out-of-core geometry. The film buffer will have a maximum size on the GPU though - independent of the render resolution.riggles wrote:V3 will be able to use system RAM for textures and geo, but I believe the frame buffer still needs to fit into the cards memory.
In theory there is no difference between theory and practice. In practice there is. - Yogi Berra
Really? My bad. I could have sworn that was in the feature list during the V3 talk at GTC.abstrax wrote:That's not correct. Version 3 will not have out-of-core geometry. The film buffer will have a maximum size on the GPU though - independent of the render resolution.riggles wrote:V3 will be able to use system RAM for textures and geo, but I believe the frame buffer still needs to fit into the cards memory.
EDIT: Also, the press release for Octane 3 should be updated, as it specifically says out-of-core geometry (under the OpenCL section). So, what's up? Did this feature get killed between then and now?
Titan X arrived today. I'm officially amazed. Threw it in the Mac with the 980 to have dual GPU power. Was a bit nervy but it started up fine, no issues. Desktop, done.
The flexibility is impressive. I can render a final on the Titan say, then open another instance of Cinama 4D, set Octane to use the 980 and not the Titan and continue working while the titan is rendering a massive print res job. Perhaps some patch renders or special passes or a totally different job.
OR I can combine both cards for faster iterative renders during production, though only up to 4GB VRAM, but that's OK because I'm only rendering tiny image buffers for tests. If the scene is big and I need a massive frame buffer I just deselect the 980 and I have access to the full 12GB again.
This is fantastic!
Edit: I just realised that I can also do work in Photoshop because the CPUs are not even being used. Ditto Mail, Safari etc. This is revolutionary for my workflow.
The flexibility is impressive. I can render a final on the Titan say, then open another instance of Cinama 4D, set Octane to use the 980 and not the Titan and continue working while the titan is rendering a massive print res job. Perhaps some patch renders or special passes or a totally different job.
OR I can combine both cards for faster iterative renders during production, though only up to 4GB VRAM, but that's OK because I'm only rendering tiny image buffers for tests. If the scene is big and I need a massive frame buffer I just deselect the 980 and I have access to the full 12GB again.
This is fantastic!
Edit: I just realised that I can also do work in Photoshop because the CPUs are not even being used. Ditto Mail, Safari etc. This is revolutionary for my workflow.
Windows 10 - 64GB RAM - Cinema 4D R20 - RTX 2070 x3
honestly - that's the best thing bout GPU render engine - especially if You have several cards on Your machine - You can render in the background without any interuptions =) & the only thing You notice is a warm breeze through Your feet =) as for PS & other programs You can still use GPU acceleration if You have discrete GPU =)sdanaher wrote:Titan X arrived today. I'm officially amazed. ...
The flexibility is impressive. I can render a final on the Titan say, then open another instance of Cinama 4D, set Octane to use the 980 and not the Titan and continue working while the titan is rendering a massive print res job. ...
OR I can combine both cards for faster iterative renders during production, though only up to 4GB VRAM, but that's OK because I'm only rendering tiny image buffers for tests. If the scene is big and I need a massive frame buffer I just deselect the 980 and I have access to the full 12GB again.
This is fantastic!
Edit: I just realised that I can also do work in Photoshop because the CPUs are not even being used. Ditto Mail, Safari etc. This is revolutionary for my workflow.

