Page 3 of 3

Re: Nvidia 3000 series support

Posted: Mon Sep 14, 2020 1:22 pm
by jimho
Timmaigh! wrote:regarding Out of Core, i have question of my own. Say you combine 24GB 3090 and 11GB 2080Ti (or 10GB 3080 or whichever GPU with smaller VRAM). Say you have a scene 18GB large. Does Octane now use 11GB of VRAM on both cards (including 3090, despite bigger VRAM) and then fills the rest with OOC, or does it use 18GB on 3090 and OOC kicks in only for the 2080Ti?
18G scene on 3 gpu(24G+2x11GvRam),18Gwill fit in 24GGPU, if Ooc is off, the two 11G GPU will fail,if ooc is on all three GPU will be working, and only the 11G GPU will use the ooc ram.

Re: Nvidia 3000 series support

Posted: Mon Sep 14, 2020 7:03 pm
by Goldorak
coilbook wrote:Hi Otoy,
Will you take advantage of rtx io where you bypass CPU and load data from ssd directly to vram https://youtu.be/QKx-eMAVK70?t=1434 ?
Yes we are working on features that leverage this - essentially making your SSD operate as ooc memory.

Re: Nvidia 3000 series support

Posted: Mon Sep 14, 2020 7:07 pm
by Goldorak
Timmaigh! wrote:And one more question, since Octane X now supports AMD hardware, should we expect regular Windows Octane with AMD support as well?
We are all looking at upcoming Ampere cards, but AMD is about release their new stuff as well, with their own hardware raytracing units, rumored massive cache and so on... for all we know it might turn out actually better performing than Nvidia...and cheaper too. If Octane happens to support it now, would be nice to know in advance.
We fully support AMD (and will continue to do so) on macOS Metal with Octane X, including raytracing HW support if that is added in the future. We also have a full version of Octane for Vulkan, but only NVIDIA drivers can run it, as AMD Vulkan tools and drivers aren't at the level of the Metal versions, so we are blocked on Octane AMD Windows until AMD can address this. When or if that will happen is uncertain.

Re: Nvidia 3000 series support

Posted: Tue Sep 15, 2020 4:12 pm
by Timmaigh!
jimho wrote:
Timmaigh! wrote:regarding Out of Core, i have question of my own. Say you combine 24GB 3090 and 11GB 2080Ti (or 10GB 3080 or whichever GPU with smaller VRAM). Say you have a scene 18GB large. Does Octane now use 11GB of VRAM on both cards (including 3090, despite bigger VRAM) and then fills the rest with OOC, or does it use 18GB on 3090 and OOC kicks in only for the 2080Ti?
18G scene on 3 gpu(24G+2x11GvRam),18Gwill fit in 24GGPU, if Ooc is off, the two 11G GPU will fail,if ooc is on all three GPU will be working, and only the 11G GPU will use the ooc ram.
Oh, thank you. I was not sure, how it works, whether the bigger VRAM card is not somehow "crippled" by the presence of the smaller VRAM card, being forced to use only aliquote part of its VRAM pool equal to smaller cards VRAM size.

Re: Nvidia 3000 series support

Posted: Tue Sep 15, 2020 4:15 pm
by Timmaigh!
Goldorak wrote:
Timmaigh! wrote:And one more question, since Octane X now supports AMD hardware, should we expect regular Windows Octane with AMD support as well?
We are all looking at upcoming Ampere cards, but AMD is about release their new stuff as well, with their own hardware raytracing units, rumored massive cache and so on... for all we know it might turn out actually better performing than Nvidia...and cheaper too. If Octane happens to support it now, would be nice to know in advance.
We fully support AMD (and will continue to do so) on macOS Metal with Octane X, including raytracing HW support if that is added in the future. We also have a full version of Octane for Vulkan, but only NVIDIA drivers can run it, as AMD Vulkan tools and drivers aren't at the level of the Metal versions, so we are blocked on Octane AMD Windows until AMD can address this. When or if that will happen is uncertain.
Thank you, this is good thing to know before any potential Ampere purchase. Anyway, does this Metal support only for AMD basically means it works on AMD hardware thanks to Apple, rather than AMD themselves?

EDIT: Leaving this here for anyone interested:

https://techgage.com/article/nvidia-gef ... rformance/

564 OB points in latest Octanebench with RTX on. Very nice. Given the CUDA core number difference, 3090 could attack 700 points, which is as much as my 2 2080Ti´s combined.

Re: Nvidia 3000 series support

Posted: Wed Sep 16, 2020 10:55 pm
by jimho
Timmaigh! wrote:
jimho wrote:
Timmaigh! wrote:regarding Out of Core, i have question of my own. Say you combine 24GB 3090 and 11GB 2080Ti (or 10GB 3080 or whichever GPU with smaller VRAM). Say you have a scene 18GB large. Does Octane now use 11GB of VRAM on both cards (including 3090, despite bigger VRAM) and then fills the rest with OOC, or does it use 18GB on 3090 and OOC kicks in only for the 2080Ti?
18G scene on 3 gpu(24G+2x11GvRam),18Gwill fit in 24GGPU, if Ooc is off, the two 11G GPU will fail,if ooc is on all three GPU will be working, and only the 11G GPU will use the ooc ram.
Oh, thank you. I was not sure, how it works, whether the bigger VRAM card is not somehow "crippled" by the presence of the smaller VRAM card, being forced to use only aliquote part of its VRAM pool equal to smaller cards VRAM size.
A scene will be fully copied into each cards' Vram seperately, so the bigger vram card will not be affected by the smaller ones, as mentioned if the scene cannot fit in the smaller one, the smaller one will fail, while the bigger one will still be working on its own(so no such cripple for the big one), this is my observation while my quadro 6000 and 2080ti work together.

probably different cards use different pools for themself, I guess