dedicated gpu server vs workstation ?

[dhoelscher]

Hello everybody,
we happen to have ten spare 1080 that we want to setup as a renderclient for octane. Now i wonder what would be the best solution to implement those cards we already own:
maybe setup 2 multigpu clients: i´d say perhaps going all x99-e ws with 5x1080 each (on air!),
or investing in a multigpu server like this one here:

https://www.mediaworkstations.net/product/gpux/

or this:

http://www.server-konfigurieren.de/konf ... -HDD-.html

though we don´t need the storage solution, since we have a dedicated 10G-Nas

we have two C4D R18 workstations with 2x titanXp each. Both Workstations should use the gpuserver or the cluster as a rendernode / octane slaves

which route would you prefer ? 2 multigpu server with octane licenses each or one big gpuserver as a slave for both workstations?

Do you have any suggestions ?

[itou31]

Hi,
I prefer to have two 5 GPUs noderender, so from one C4D you can connect to one or 2 rendernodes. If two c4d need to render simulteneously, you can connect one rendernode to each. I don't know if you can dispatch gpu on slave node for two master nodes.