Quad Gigabyte AIO 2080-Ti, dual NVLINK ...

[JinnGonQui]

Quad Gigabyte AIO 2080-Ti (GV-N208TAORUSX-W-11GC), dual NVLINK  ==  quad'ish performance, double'ish VRAM ??? ... for Arch Viz in Unreal Engine 4 (Unreal Studio / DataSmith) ...

.

... so, following the thought process:
if a software supports NVLINK -- AND it supports memory pooling, being only across the NVLINK, 

the benefit to Arch Viz would be :

• maximizing VRAM for GPU Lightmass baking which is massively VRAM hungry,
• as well as "should" take advantage of all available GPU's in box ....

.

... so the question is :

Does Unreal Engine 4 (Unreal Studio) support memory pooling via NVLINK ?

. 

. 

references :

https://www.pugetsystems.com/labs/articles/NVLink-on-NVIDIA-GeForce-RTX-2080-2080-Ti-in-Windows-10-1253/#DoGeForceRTXCardsSupportMemoryPoolinginWindows?

https://forums.unrealengine.com/development-discussion/rendering/1460002-luoshuang-s-gpulightmass?p=1600770#post1600770

https://www.nvidia.com/en-us/titan/titan-rtx/#textcomponentenhance593c4762_06f5_4894_a6a1_3f7a26e70a60

https://www.newegg.com/gigabyte-geforce-rtx-2080-ti-gv-n208taorusx-w-11gc/p/N82E16814932075?item=N82E16814932075&source=region&nm_mc=knc-googleadwords-pc&cm_mmc=knc-googleadwords-pc-_-pla-_-video+card+-+nvidia-_-N82E16814932075&gclid=CjwKCAjw7uPqBRBlEiwAYDsr14lZlurSCWoQyoG7OuorWkFVjZss_6vKwkbIfLwyoedfmqKDZZu5HxoCQLUQAvD_BwE&gclsrc=aw.ds

[cried22]

Memory pooling is not supported on the 2080ti to my understanding. That is only a workstation card feature.

[TrigrH]

13 minutes ago, cried22 said:

Memory pooling is not supported on the 2080ti to my understanding. That is only a workstation card feature.

the puget link above seems to think otherwise

[astranger200]

40 minutes ago, JinnGonQui said:

Quad Gigabyte AIO 2080-Ti (GV-N208TAORUSX-W-11GC), dual NVLINK  ==  quad'ish performance, double'ish VRAM ??? ... for Arch Viz in Unreal Engine 4 (Unreal Studio / DataSmith) ...

.

... so, following the thought process:
if a software supports NVLINK -- AND it supports memory pooling, being only across the NVLINK, 

the benefit to Arch Viz would be :

• maximizing VRAM for GPU Lightmass baking which is massively VRAM hungry,
• as well as "should" take advantage of all available GPU's in box ....

.

... so the question is :

Does Unreal Engine 4 (Unreal Studio) support memory pooling via NVLINK ?

. 

. 

references :

https://www.pugetsystems.com/labs/articles/NVLink-on-NVIDIA-GeForce-RTX-2080-2080-Ti-in-Windows-10-1253/#DoGeForceRTXCardsSupportMemoryPoolinginWindows?

https://forums.unrealengine.com/development-discussion/rendering/1460002-luoshuang-s-gpulightmass?p=1600770#post1600770

https://www.nvidia.com/en-us/titan/titan-rtx/#textcomponentenhance593c4762_06f5_4894_a6a1_3f7a26e70a60

https://www.newegg.com/gigabyte-geforce-rtx-2080-ti-gv-n208taorusx-w-11gc/p/N82E16814932075?item=N82E16814932075&source=region&nm_mc=knc-googleadwords-pc&cm_mmc=knc-googleadwords-pc-_-pla-_-video+card+-+nvidia-_-N82E16814932075&gclid=CjwKCAjw7uPqBRBlEiwAYDsr14lZlurSCWoQyoG7OuorWkFVjZss_6vKwkbIfLwyoedfmqKDZZu5HxoCQLUQAvD_BwE&gclsrc=aw.ds

yeah no. 2080 tis dont memory pool. thats a quadro only option

[JinnGonQui]

100% clear from the horses mouth on the hardware capability (https://www.nvidia.com/en-us/deep-learning-ai/products/titan-rtx/) ... looking for confirmation on the engine side of things