Interesting Read: 6870 vs 560 (non ti)

What opened your eyes? I don't get it.

No matter what card, some games play better on ATI cards and some play better on Nividia cards, they have 2 different architectures.

It's always been that way, no new news here.

Nvidia released the 560 to counter the 6870. AMD released the 6950 1GB to counter the GTX560ti.
 

precursoris

Distinguished
Mar 7, 2010
152
0
18,690


Being new to gaming builds, I was under the impression that frames per second was the primary aspect when comparing GPU's. This review though points out that we need to look at the frame time latency to get a more in-depth feel of how a card performs relative to another.

The problem with FPS over time is that FPS only looks at one second at a time. If a game is running at 60 fps, that means it is taking 60 points of data, and averaging them out over a second. That doesn't sound like much, but in the 'feel' of a game, one second is a very long time. This method looks at every single frame that is passed through the renderer, and times them. That's why Witcher 2 can run at around the same average FPS on both cards, but much worse on the 6870. Some frames are being delivered at <10 "fps" while others are dropped completely to keep up. That's why you get micro stuttering.
 

cuecuemore

Distinguished
Instantaneous FPS will always be the reciprocal of frame time. For this reason people look at the minimum FPS (highest single frame time) as a gauge of playability rather than average FPS. Their FPS graphs are shown on a per-second average basis, which provides the illusion of high framerates.
 

Desert Eagle

Honorable
Mar 26, 2012
107
0
10,710


True, and minimum FPS usually comes when there's the most action on screen. For instance in FP Shooters when the enemy is throwing everything at you and you don't need the framerate drop or stuttering the most. I wish more tech sites benchmarked minimum FPS. That would be more useful info than average FPS.