$1200 GPU set up, 2014 edition Q

[Broc Kelley]

Q: for 3 monitor set up, 3240x1920 @60hz, what is the best dual gpu set up? Answer what you want for any specific build, but my build has an fx 8350 cooled by a Noctua NH-D14; I play all kinds of games but have mostly been playing Skyrim and BF4, along with slight bouts of free to play MMOs and Planet Side 2. Again, I'm on air and I'd like to continue to be on air with my graphics cards.

I know the non reference 290s just came out, and I know the 790 and Titan Black edition are on there way.I also know I can OC any 780 and reach 780 ti base-clock performance, and I know 780ti's are on par with r9 290x for the most part. I know that Mantle finally went live for the public and people are saying it actually works, and I know that 3gb of VRAM might not be enough for the 780s because of how capable those cards are..The problem is in the pricing. As a college student I really shouldn't shift that budget around too much..it sucks eating ramen for a month to get the latest and greatest..but hell, it's why we're here. Any help?

My top choices:

Windforce r92 290~$625
XFX DD r9 290~$600
Classified 780~$575
MSI Lightning 780/ ACX 780~$550
Windforce 780~$500

I'd SLI these obviously.

The questions that come to mind:
Will the VRAM make a difference?

Will the Gaming industry change so much that these cards don't matter as much as the servers we run games on?

Should I/We plan for the gaming industry to change in this fashion, or should we continue to invest fully?

Should I even be investing $1200 when all this APU stuff is going on?

As the best price/performance is obviously the windforce gtx 780 at $500, is the +1gb of VRAM and current Mantle support worth the extra $100 given that next gen consoles are running AMD stuff? or since console game's are hardware-optimized does that not make any difference?

 

[vince232]

if you're going for 3 monitor setup go with the xfx r9 290 and crossfire it. amd has better drivers for the tri monitor display.

here mate hope this helps you

http://www.youtube.com/watch?v=N5fi3ARXWKc

it is only a 7 minute video please watch it

 

[houldendub]

if you're going for 3 monitor setup go with the xfx r9 290 and crossfire it. amd has better drivers for the tri monitor display.

here mate hope this helps you

http://www.youtube.com/watch?v=N5fi3ARXWKc

it is only a 7 minute video please watch it

Did AMD ever fix frame pacing on eyefinity setups?

 

[Broc Kelley]

if you're going for 3 monitor setup go with the xfx r9 290 and crossfire it. amd has better drivers for the tri monitor display.

here mate hope this helps you

http://www.youtube.com/watch?v=N5fi3ARXWKc

it is only a 7 minute video please watch it

Is the only reason you choose the DD 290 because of drivers?

Question: Could I automatically assume that one screen at 4k (8,000,000 pixels) will be harder to drive than 3 screens each at 1920x1080 (6.2 million pixels) in direct relation to the resolution differences, or does the fact that they are 3 separate screens cause a supplementary detriment to performance? If you're answer is in any way, " the fact that it's 4k hits harder than the fact that it's 3 screens" could we then justify higher 4k results one gpu vs another, automatically equals higher 3240x1920 results for that same GPU? Or is it equally crazy to assume that proof as to assume that if one gpu is better at 2560x1600 it will be better with 2 million more pixels as well?

2560x1600 is 2M pixels less than 3240x1920 which is 2M pixels less than 3840x2160

Thanks for the reply. I've watched the video before, and when I read forums I do see people coming to the same answer as you: AMD, better 3 screen drivers. But,when I look at the actual GPU reviews and benchmarks, besides almost always seeing whatever gpu they're reviewing getting the highest scores, it really seems even. The 780ti matches well with the r9 290x with the 290x's getting better scores at base clock, and the 780ti having better overclocking, noise, heat, and power consumption. So, yes it is easy to come to the conclusion that if all I want is performance at base clocks, ceteris paribus, 290 dual dissipation it is for now with these particular drivers. But with the gpu market being so dynamic, that really isn't the only factor.

If you could please, extrapolate and maybe link actual benchmarks showing the most recent drivers for the GPU's in question. I've looked all over at Hardocp, guru3d, hardwarecanucks, here at Tom's, etc etc.. and like I said it depends not only what test-setup they're using but if that test setup in particular is a golden egg so to speak. is the answer really as LInus proclaims, luck determines how well your system will run?..I suppose that question is equally useless to answer as it is to have as an answer..

 

[Broc Kelley]

Did AMD ever fix frame pacing on eyefinity setups?

Edit: not sure if we can post links here buuut, hothardware.com/News/Imminent-AMD-Catalyst-Driver-Release-To-Enable-Mantle-Fix-Frame-Pacing-Issues-and-Support-HSA-Features-on-Kaveri/ apparently the beta drivers help with frame pacing, though the higher the resolution the smaller the percentage bump..which is bad since the higher the resoution the lower the fps, that means an absolute AND a marginal disparity for those above "1600p"

Back in December a couple articles surfaced about them doing something about it in January, so I'm guessing... not. They said they were cranking out Drivers once a week complimentary to their continuance of Mantle proprieties, so maybe they released both together...though reviews of Mantle with BF4 have suspiciously only included 3 instances all of which run on only 1 screen.. to my knowledge anyway

 

[vince232]

if you're going for 3 monitor setup go with the xfx r9 290 and crossfire it. amd has better drivers for the tri monitor display.

here mate hope this helps you

http://www.youtube.com/watch?v=N5fi3ARXWKc

it is only a 7 minute video please watch it

Is the only reason you choose the DD 290 because of drivers?

Question: Could I automatically assume that one screen at 4k (8,000,000 pixels) will be harder to drive than 3 screens each at 1920x1080 (6.2 million pixels) in direct relation to the resolution differences, or does the fact that they are 3 separate screens cause a supplementary detriment to performance? If you're answer is in any way, " the fact that it's 4k hits harder than the fact that it's 3 screens" could we then justify higher 4k results one gpu vs another, automatically equals higher 3240x1920 results for that same GPU? Or is it equally crazy to assume that proof as to assume that if one gpu is better at 2560x1600 it will be better with 2 million more pixels as well?

2560x1600 is 2M pixels less than 3240x1920 which is 2M pixels less than 3840x2160

Thanks for the reply. I've watched the video before, and when I read forums I do see people coming to the same answer as you: AMD, better 3 screen drivers. But,when I look at the actual GPU reviews and benchmarks, besides almost always seeing whatever gpu they're reviewing getting the highest scores, it really seems even. The 780ti matches well with the r9 290x with the 290x's getting better scores at base clock, and the 780ti having better overclocking, noise, heat, and power consumption. So, yes it is easy to come to the conclusion that if all I want is performance at base clocks, ceteris paribus, 290 dual dissipation it is for now with these particular drivers. But with the gpu market being so dynamic, that really isn't the only factor.

If you could please, extrapolate and maybe link actual benchmarks showing the most recent drivers for the GPU's in question. I've looked all over at Hardocp, guru3d, hardwarecanucks, here at Tom's, etc etc.. and like I said it depends not only what test-setup they're using but if that test setup in particular is a golden egg so to speak. is the answer really as LInus proclaims, luck determines how well your system will run?..I suppose that question is equally useless to answer as it is to have as an answer..

coming from what you said. r9 290x matches 780ti well. even better at base clocks, and the 780ti is better when overclock comes to the picture, noise, temp etc.

considering how much the 780ti is, i think r9 would be more bang for his buck

 

[coozie7]

Make your mind up what display/s you're going to use because for gaming it's the total resolution that counts, the more pixels the card has to render the higher the workload, regardless of the number of screens or the overhead created by CF/SLI.
Using remote servers/renderers is a nice idea but a long way off, the cost of building and running such a system would be huge and even if its output were restricted to normal HD there is still plenty of regions in the world where the infrastructure would need to be upgraded to cope. We'll be gaming on our own rigs for some time to come.