• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

7970 oc or 670?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
vRAM is mirrored (if that is what you meant by shared), correct. So you only have 2GB with 2 2GB cards as the same data is on both cards. It is not a pool (4GB available).

If I understand what you are saying... when you add more cards you are not increasing 'core capabilities and with higher core clocks on all cards'. In fact, it is quite common to NOT be able to overclock as much with a single card vs 2 or more. If more ram isnt required, more ram will not be used in SLI/Crossfire or overclocking for the same amount of pixels needing rendered.

... or did I misunderstand what you are saying? LOL!
 
vRAM is mirrored (if that is what you meant by shared), correct. So you only have 2GB with 2 2GB cards as the same data is on both cards. It is not a pool (4GB available).

If I understand what you are saying... when you add more cards you are not increasing 'core capabilities and with higher core clocks on all cards'. In fact, it is quite common to NOT be able to overclock as much with a single card vs 2 or more. If more ram isnt required, more ram will not be used in SLI/Crossfire or overclocking for the same amount of pixels needing rendered.

... or did I misunderstand what you are saying? LOL!

Sort of.

I'm saying since basically by adding more cards you effectively get more core power but you have the same amount of framebuffer as the single card. So with the increased power of having 3 cards but having a comparatively small amount of VRAM, you could run into bottlenecks in the VRAM department. And we all know how unplayable running out of VRAM makes a game.
 
Last edited:
Gotcha. But... It doesnt matter if I have 1/2/3/4 cards the game will use the same amount of memory. Adding more cards and having more GPU power does not necessitate more ram or ram bandwidth.

Here is how SLI works. Hopefully having that understanding can help:
One is called split frame rendering - this is where the frame it split horizontally in half, which splits the workload 50/50 for each card. It intelligently splits the workload based on the geometry of the frame, for example if there is little to render on the top half (e.g. sky) and a lot on the bottom (smoke, debris, ground textures) the dividing line will move to balance the workload between each card. Split frame rendering is pretty poor at scaling the work when compared against alternate frame rendering (below).

The second is called alternate frame rendering. Each card renders full frames, but in sequential order, for example the master card might process even frames (2,4,6) while the slave card processes odd frames (3,5,7). When the slave card finishes work on a frame the results are sent to the master card, usually via the SLI bride (as described above). Theoretically this should result in the rendering time being cut in half, which should technically linearly scale the performance by double. This is not often the case in practice though, depending on drivers, games, etc.
http://www.overclock.net/t/1210604/how-does-sli-work
 
Last edited:
You have the same effective vram size (it hasn't been just a framebuffer in a while), but that's just fine as you have the same monitor size as well. You don't need more vram with more GPUs.

Detail level and resolution are what matter vram wise.
It could be argued that multiple cards allows setting the detail higher, that's all well and good, but as far as I know it isn't an issue yet, even maxed out on a single large monitor.
 
Yeah but what I mean is that if you add more GPUs then if you use more VRAM (from multiple monitors or high resolutions) then you have the GPU power but not the VRAM size.

So it can become a bottleneck, as I said before. Having more VRAM eliminates this problem.
 
the 670/680 are more game orientated and out of the box performance is very good go that route.
 
Yeah but what I mean is that if you add more GPUs then if you use more VRAM (from multiple monitors or high resolutions) then you have the GPU power but not the VRAM size.

So it can become a bottleneck, as I said before. Having more VRAM eliminates this problem.
you never mentioned, until now, multi monitors. You said 25x14/1600. With multi monitors, as I stated way earlier, you would be correct.

Hard to know what u r talking about when its not mentioned and different than the established context...
 
you never mentioned, until now, multi monitors. You said 25x14/1600. With multi monitors, as I stated way earlier, you would be correct.

Hard to know what u r talking about when its not mentioned and different than the established context...

Oh come on, that's a low blow.

2560x1600 is a big resolution, especially in games like BF3 that use upwards of 1200mb at 1080p. I'd imagine at triple 1080p then it would be ridiculously high.

I'm still undecided on the monitor setup, but it will either be 30" 2560x1600 or triple 1920x1080. Leaning towards a single 30". But that's down the line a couple months.
 
Low blow? Not sure what would be considered a low blow. You were basing your posts on a talking point you did not share (the possibility of multi monitor). Im simply saying it helps to have all that information up front.

That said, triple monitor, though now it isnt showing to be a problem, I would still get more than 2GB at that point.

GL. :)
 
Back