[Various] Dual Fiji revealed

[3DVagabond]

It's not backwards. Not sure what some of you are seeing?

 

[3DVagabond]

There's still a limit on 4 gpus for CF X

That's DX limit. Although, to be honest, I can't see them wanting to invest in additional driver work for more GPU's even if it was possible.

 

[sm625]

If DX12 works as well as they say, AMD should just go ahead and squeeze 3 of these onto one board.

 

[LTC8K6]

If DX12 works as well as they say, AMD should just go ahead and squeeze 3 of these onto one board.

825W on one board?

 

[Silverforce11]

Better to put Fury Nano x 3 or x4 on one board, 2x the efficiency is better than 1.5x

 

[Headfoot]

Given the push towards job based systems, and the fact that GPU processing is already massively parallel, it should be theoretically possible to scale to many GPUs, well over 4. The limitation will likely be related to how much overhead (both actual processing time and developer man hours) it takes to split the tasks to such a small level of granularity and distribute them. My guess is only the big engines are going to have explicit multi adapter, less of those will have iGPU supporting multi adapter, and even less than those (perhaps just one or two) will have multi adapter support exceeding 4 GPUs.

 

[DownTheSky]

If one Fiji GPU can be under 175W as Fury Nano and still be over 290X perf, I'm sure they can make a dual-gpu with even under 350W TDP and still have crazy perf. Though knowing AMD they'll go for the 500W version.

Would be cool to have something like 350W mode bios switch.

 

[exar333]

If this uses < or ~500w, it might be worth waiting for (skipping 980Ti and current Fury). It would be killer for the upcoming Oculus Rift and would have plenty of grunt to run a newer game at max quality and 90fps (OR is 1200x1080 (x2) displays @ 90hz) so similar to a 1440P display at 60hz for driving.

With my current loop and PSU, this would be a perfect drop-in and save me some effort.

Decisions, decisions.