- Joined
- 14 Dec 2010
- Posts
- 495
Back on topic...
Very, very interesting article here. It appears that memory bandwidth itself is the limiting factor at very high resolutions/multi monitor setups, where the gap between the GTX 680 and the 7970 all but diminishes, particularly in highly overclocked scenarios.
I'm still waiting to see what the benches are on the 4GB cards, but if bus bandwidth is the limiting factor, it seems tacking an extra 2GB of VRAM onto it is kind of insignificant if there's a bottleneck using it, right?
Very, very interesting article here. It appears that memory bandwidth itself is the limiting factor at very high resolutions/multi monitor setups, where the gap between the GTX 680 and the 7970 all but diminishes, particularly in highly overclocked scenarios.
I'm still waiting to see what the benches are on the 4GB cards, but if bus bandwidth is the limiting factor, it seems tacking an extra 2GB of VRAM onto it is kind of insignificant if there's a bottleneck using it, right?