OTOY Forums

Hi all,

I'm new to this so feel free to correct me, I'm here for guidance!
I'm interested in the idea of using PCIe risers to add more GPUs than you could fit in a normal case/mobo with cards in slots next to each other (limit of 4 x double-width cards)
The plan would be to have 1 card in the case for the C4D UI, and then 4-6 outside for rendering. However, I know the PCIe interface then becomes only x1 instead of x16.

Here are my 2 Qs:

Rendering
Am I right in thinking for final rendering in C4D the actual speed between a card on a riser and a card in an x16 slot wouldn't be any different? - Just the time it takes to load the scene into the card would take longer over the riser, but once it's on the card the speed is all good? Are we talking a long wait? And what about sending data back form the card?

Viewport performance
Here's my big concern - will there now be a delay between the C4D viewport and the Octane window? Will it feel less responsive? Playing the timeline in C4D will there be a delay between the viewport animation and the Octane window?


Thanks so much to anyone who can answer these 2 Qs

No for both questions..

Only speed drop is during loading phase - everything else is same, render speed and viewport update is normal
- it's the same with games, when the game is loaded it's all good no lag whatever, the scene is on the gpu vram now

Hi acc24exm thanks for your answer!

So does Octane work by caching the scene to all the cards when the project is opened? What happens if you create new objects in the scene, do these then get cached automatically?

Apologies as I don't have a computer that can run Octane to try this stuff out myself.