Hi Peter
Oh, that is interesting. I didn't know that the slowest card sets the speed limit. Are we purely talking about the MHz? Is that a OSX or Resolve thingy?
So are there any whitepaper on what the GUI-gpu is used for or so? I like stuffs like this and really want to dig my teeth into it to get a better understanding. When doing tests, maybe not the most scientifically tests, the GUI-GPU doesn't seem to be that important. I have a spreadsheet with my tests over here:
https://docs.google.com/spreadsheet/ccc ... TZnc#gid=0 When I look at that it feels that in a real life scenario the user would gain more from being able to have the GPU GUI used for video processing too. I know that I don't see the whole picture so I know that I'm on thin ice and maybe in over my head. =)
But for instance compare the numbers for this setup:
Video card 1: GeForce GTX 580 - GUI Only
Video card 2: GeForce GTX 690
Video card 3: GeForce GTX 690
Video card 4: GeForce GTX 680
With the numbers for this setup where there's an ATI card in the MacPro:
Video card 1: GeForce GTX 580
Video card 2: GeForce GTX 690
Video card 3: GeForce GTX 690
The later setup gives pretty much identical results in Resolve but if the 580 in the first setup could be used for more than GUI I'm certain that it would see a pretty big boost in benchmarking. Also, why isn't the ATI reported as a GUI only GPU? When running a system with only an ATI it is reported in the lists as "Video card 1: ATI Radeon HD 5770".
I admit I might be a bit snowed in on numbers but that's what testing and tweaking does to ya.
Cheers