My issue I seem to be having on one system is NVlink is plugged in, SLI States enabled maximize 3D, but still the information is duplicated. both cards are 2080's and the scene is a total of 11.6 gb vram. on CUDA ,system crashes or blender crashes, or goes right to launch error Cuda memory, on Optix, render gets half way through then provides CUDA error. any solution on a fix? or is the quote below false and we cannot free up the extra memory on 2 cards. Next does this apply to 10 series cards in SLI, all my cards have SLI plugged in now for testing I have x3 Titan XP's on one system and all have duplicated 11.6 gb vram.

System in Question

core i7 8700K 64gb Ram x 2 MSI 2080's

Blender website Quote

"Allocates resources across multiple GPUs rather than duplicating data, effectively freeing up space for larger scenes. Note that in order for this option to be available, the GPUs must be connected together with a high bandwidth communication protocol. Currently only NV Link on Nvidia GPUs is supported."

Any Insight would be Great :)

  • $\begingroup$ I was curious about this myself. I found this article: pugetsystems.com/labs/articles/… Appears memory pooling is only available on Quadro platform. SLI is unfortunately not very useful on the consumer side anymore. $\endgroup$ Nov 20, 2020 at 0:00


You must log in to answer this question.

Browse other questions tagged .