Hi all. First time poster so I'm not sure if this is the best place to be starting this topic?
I recently used PCPartPicker to build a new work machine. A lot of my work involves Cinema 4D, X-Particles (physics simulations) and obviously Octane, so I built it with that in mind. However, I also had an idea of expanding it to a more powerful, multiple GPU render machine. And that's where I'm hoping you clever lot can help me out.
My current spec:
Fractal Design Define R6 Black TG Midi Tower Black
Asus Prime Z370-A
8th Gen Core i7-8700K
4 x Corsair 16GB DDR4 Vengeance LPX 3000MHz
1x GTX 1080ti - 11GB GDDR5X
1 x 1TB Samsung 970 EVO M.2 NVMe SSD
PSU: Corsair CP-9020094-UK RM1000
I am considering buying another GTX 1080ti and I also have a GTX 1080 to throw into the mix.
The first question I have is regarding the best way to configure any additional graphic cards, and avoid bottle-necking with things like PCIE bus speeds? I'm aware that the motherboard will run:
1 x PCIE at 16x
2 x PCIE at 8x
1 x PCIE at 8x plus 2 x PCIE at 4x
What I'm less aware of and have found difficult to find an answer on line, is what GPU cards are bottle-necked by what PCIE bus speeds? And how that might effect performance for GPU rendering? Also, if I want to go down the route of adding GPU cards with a PCIE riser, what limitations in terms of bottle-necking come in to play then? There appears to be a lot more detailed information on bit mining rigs, and whilst they are similar setups, I also gather that they have different performance requirements.
Any help will be massively appreciated, Thanks, Jamie