Hi there
I'm using a newly installed system:
MB Gigabyte B550 Vision D-P
AMD 5900X 12 core CPU
64GB DDR4 memory
two 2TB nvme SSDs
Two EVGA 3080ti with 12GB Vram
Installed with windows 10 pro
running a desktop resolution of 2560 x 1440 connected via HDMI 2.1 cable to GPU1
i'm aware that currently (without using nvlink - not sure not even then) the memory
used by octane is just from a single GPU.
As attached image shows almost 4GB of memory (out of 12GB) is UNAVAILABLE. I really want to understand why not almost all the GPU Vram is available
to Octane. running just the OS and almost no other app other than 3dsmax itself should not take that much
GB of Vram. what am i missing? do i have an issue or this looks ok to you?
cheers
GPU memory / management techniques?
Forum rules
Please post only in English in this subforum. For alternate language discussion please go here http://render.otoy.com/forum/viewforum.php?f=18
Please post only in English in this subforum. For alternate language discussion please go here http://render.otoy.com/forum/viewforum.php?f=18
-
paride4331
- Posts: 3813
- Joined: Fri Sep 18, 2015 7:19 am
Hi MartinBoom,
In multi-GPU setups, the amount of RAM available to OctaneRender is not equal to the sum of the VRAM on the GPUs, but it is restricted to the GPU with the smallest amount of VRAM on board (in your case 12GB). Also keep in mind that many applications use VRAM (your 3D application, editors, internet browser ..) The monitor also needs VRAM. also Windows 10 (Windows 11 a little less) eats around 18% VRAM. Either way Out of core will help a lot when a VRAM limit. At least 8GB.
Regards
Paride
In multi-GPU setups, the amount of RAM available to OctaneRender is not equal to the sum of the VRAM on the GPUs, but it is restricted to the GPU with the smallest amount of VRAM on board (in your case 12GB). Also keep in mind that many applications use VRAM (your 3D application, editors, internet browser ..) The monitor also needs VRAM. also Windows 10 (Windows 11 a little less) eats around 18% VRAM. Either way Out of core will help a lot when a VRAM limit. At least 8GB.
Regards
Paride
Last edited by paride4331 on Wed Sep 28, 2022 7:46 am, edited 2 times in total.
2 x Evga Titan X Hybrid / 3 x Evga RTX 2070 super Hybrid
-
MartinBoom
- Posts: 59
- Joined: Mon Apr 02, 2018 8:10 pm
hi
thanks all for the reply.
Paride, i knew that the Vram doesnt add up as you add more GPUS.. my question goes if there is anyway to free-release that `unavailable` Vram specially when one only renders with every little app closed.
cheers
thanks all for the reply.
Paride, i knew that the Vram doesnt add up as you add more GPUS.. my question goes if there is anyway to free-release that `unavailable` Vram specially when one only renders with every little app closed.
cheers
-
MartinBoom
- Posts: 59
- Joined: Mon Apr 02, 2018 8:10 pm
and thanks Eslksa, allways helpful your articles/posts.
my guess is that it should be software that allows you to give the memory of the GPU priority. currentely rendering and the unavailable GPU Vram is almost 5GB. out 12!..ridiculous
my guess is that it should be software that allows you to give the memory of the GPU priority. currentely rendering and the unavailable GPU Vram is almost 5GB. out 12!..ridiculous
This problem should be solved two years ago. It was a very long term discussion: viewtopic.php?f=12&t=51992&hilit=win10+vram&start=110
I use out of core rendering a lot. It was fairly unstable with my old system. But since my hardware upgrade in April out of core rendering works very well. One time stable with 11GB out of core.
Maybe it's time for an upgrade or try Windows 11.
I use out of core rendering a lot. It was fairly unstable with my old system. But since my hardware upgrade in April out of core rendering works very well. One time stable with 11GB out of core.
Maybe it's time for an upgrade or try Windows 11.
PURE3D Visualisierungen
Sys: Intel Core i9-12900K, 128GB RAM, 2x 4090 RTX, Windows 11 Pro x64, 3ds Max 2024.2
Sys: Intel Core i9-12900K, 128GB RAM, 2x 4090 RTX, Windows 11 Pro x64, 3ds Max 2024.2