Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
There's a pattern in those severely GPU limited tests that I've seen. The 3x efb scale fps score is always halved of the 2x efb scale on those GPUs limited test. However, this does not happen to the mid-range GPUs like the GTX 260, GTX 460, HD 5850, and so on. So, if anyone is going to buy a new graphic card soon or in the coming future, don't go for the ones that produced half of the 2x efb scale fps at 3x efb scale.
For example:
1x efb scale 89 fps
2x efb scale 33 fps
3x efb scale 16 fps
But there's an exception to this--check this out:
Core i7 970 @4.2GHZ ,GTX460 1 GB default
1x efb scale 302 fps
2x efb scale 261 fps
3x efb scale 136 fps
Even though at 3x efb scale, the fps is almost half of 2x efb scale, but look at the fps. That's a whole different story! To me, the most acceptable fps at 3x efb scale is 60 fps, recommended is 80-90+ fps.
Quote:There's a pattern in those severely GPU limited tests that I've seen. The 3x efb scale fps score is always halved of the 2x efb scale on those GPUs limited test. However, this does not happen to the mid-range GPUs like the GTX 260, GTX 460, HD 5850, and so on. So, if anyone is going to buy a new graphic card soon or in the coming future, don't go for the ones that produced half of the 2x efb scale fps at 3x efb scale.
This is because a high internal resolution increases the chance of a linear performance relationship due to the increasing amounts of throughput needed for per-pixel work such as pixel shading. And since 3x efb scale is just over twice the resolution of a 2x scale.....
1x efb scale: .33MP (640 x 528)
2x efb scale: 1.33MP (double width double height = 4x resolution)
3x efb scale: 3.0MP (triple width triple height = 9x resolution of native)
The point is, you don't want to buy a graphic card that can't produce 60+ fps (minimum) 3x efb scale. Also, are all Wii or Gamecube games run at 640 x 528 internal?
Intel Xeon E5462 @ 2.8 GHz with ATI Radeon HD 4870, 700 MHz core, 1090 MHz memory (MacPro)
Screen res: 1920x1200
1x efb scale 125 fps
2x efb scale 122 fps
3x efb scale 81 fps
3x efb scale and SSAA 4x 22 fps
Quote:Also, are all Wii or Gamecube games run at 640 x 528 internal?
No. Widescreen games have a different internal resolution, 720 x 528.
Quote:with ATI Radeon HD 4800,
Please be more specific. 4830? 4850? 4870? 4890?
(01-15-2011, 01:20 PM)NaturalViolence Wrote: [ -> ]Quote:Also, are all Wii or Gamecube games run at 640 x 528 internal?
No. Widescreen games have a different internal resolution, 720 x 528.
Quote:with ATI Radeon HD 4800,
Please be more specific. 4830? 4850? 4870? 4890?
Sorry. Post edited. It's an ATI Radeon HD 4870
I am sorry for the lack of updates in the first post but the participation became bigger than i expected. There is probably some need for a better organization of the data and maybe a different type of graph since the current one is congested and will become a mess if data continue to be added, along with a graph that takes into account anti aliasing. Haven't quite made my mind for it so be patient for a while.
The data graph along with the formating of the post is updated. What is left is taking into account anti aliasing in a second graph and the antialiasing data be included in the excel file. The new graph is more flexible and can add as many data as possible.
the think i saw in graph correct me if i make a mistake but more the resolution efb scale increase the impact of cpu like in pc games when i saw benchmark in low resolution intel was so good and more the level of resolution less the different was saw
it hard to compared test and like i am nearly the only one with amd and i dont saw anybody with intel and 470 is hard to really see impact
Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19