I'm looking into purchasing a license for Modo, but I'm noticing something strange when I render various scenes.
When I run a render with a single GTX 970, my used VRAM and my free VRAM add up to a logical number, somewhere around 3.5-3.8 gigs, but if I enable my second GTX 970, my free VRAM drops dramatically. Basically, it seems to only recognize 2.5gb of my second 970, and I can't determine if this is an Octane issue, or an issue with my card. It'll say 4gb total, but only 2.5 or so is free, even on low poly scenes. Games and other things seem to recognize the full 4gb as far as I can tell, drivers are also updated.
It seems to only happen if I enable my secondary card, or enable it solo, and it seems to happen in the standalone renderer as well.
Hopefully someone can help, thanks in advance!
Free VRAM reduced with two GPUs
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.
It actually doesn't seem to be consistent with the standalone, I checked again and it seems fine.
However, I noticed on a 13 million poly scene in Modo, that's when it gets the 2.5gb cap. Then, if I go back to another scene, THAT'S when it'll start capping that one. If I close Modo and re-open it and open a smaller scene, the VRAM seems fine. So for some reason, a really high polygon scene seems to take a toll on the VRAM.
I'm speculating that it's because Modo needs to load it into the OpenGL? But it still seems odd to me that it's specific to the second card.
However, I noticed on a 13 million poly scene in Modo, that's when it gets the 2.5gb cap. Then, if I go back to another scene, THAT'S when it'll start capping that one. If I close Modo and re-open it and open a smaller scene, the VRAM seems fine. So for some reason, a really high polygon scene seems to take a toll on the VRAM.
I'm speculating that it's because Modo needs to load it into the OpenGL? But it still seems odd to me that it's specific to the second card.
After a little more digging into it and using EVGA's Precision tools, it appears that the SECOND card listed in the CUDA settings is actually my primary card. So someone please correct me if I'm wrong, but it appears everything is simply capping to the amount free it has after the main card handles openGL. And the reason I get full VRAM using JUST the secondary card to render (originally thought to be my main card) is because the main card is handling all the openGL.
It doesn't quite account for the inconsistent amounts it will sometimes use for the same scene, but that sounds like Modo's issue in how it uses and flushes the VRAM.
It doesn't quite account for the inconsistent amounts it will sometimes use for the same scene, but that sounds like Modo's issue in how it uses and flushes the VRAM.