Question for people with beastly gaming rigs:

[Chalenge]

I think that refers to actual supported monitor refresh, not the fps you can achieve vsync off.

Running vsync off is useful as means of benchmark because it can tell you how much higher possible performance your graphics card has at the given load. If fps would be pegged at 60 you wouldn't see any difference between 100% and 300% hardware performance unless you had a benchmark that can put even the 300% hardware to its knees.  I don't think ramdac has anything to do with vsync off fps rates because the output never goes that far. AFAIK Ramdac handles only the output that's actually sent to the monitor.

I do not see it as useful, since the only test you are actually performing is a memory buffer write test. Since the data is not actually passing through the RAMDAC it is merely a record of the number of times the memory buffer write has occurred, which is only a portion of the process. You just said it yourself after all.

So, the readings beyond actual monitor refresh and card refresh rates are pure fiction (fantasy) as I said.

[MrRiplEy[H]]

I do not see it as useful, since the only test you are actually performing is a memory buffer write test. Since the data is not actually passing through the RAMDAC it is merely a record of the number of times the memory buffer write has occurred, which is only a portion of the process. You just said it yourself after all.

So, the readings beyond actual monitor refresh and card refresh rates are pure fiction (fantasy) as I said.

Lol no its not. If you artificially tie your performance to a limit you can never test the theoretical max performance. At last not without finding some tasks for the GPU that stress it far enough to drop under those limits.

So you're saying that if two people benchmark their computers qith Quake4 and the other gets 45fps while the other gets 400fps, that's all theoretical and has no value? Basically you're saying that benchmarking has no value then. You just killed 3DMarks industry lol.

[Chalenge]

All I am saying is that there is a difference between pre-RAMDAC, and post-RAMDAC refresh. All this test is doing is testing the memory page refresh. It is not a good benchmark of anything more than that. It is not testing the complete process. Sorry you can't see that.

[MrRiplEy[H]]

All I am saying is that there is a difference between pre-RAMDAC, and post-RAMDAC refresh. All this test is doing is testing the memory page refresh. It is not a good benchmark of anything more than that. It is not testing the complete process. Sorry you can't see that.

Its not testing memory only, it tests GPU performance, shader performance etc. Ramdac is not a bottleneck unless it's made horribly wrong.

[Chalenge]

Well, that is until you realize that there can be quite a difference between internal benchmark and what is actually ever seen on screen. That number can tell you something along the lines of how everything pre-RAMDAC (or today it might be TMDS) turns out, but there is still a lot more beyond that depending on the monitor(s) you are using. For instance, I get 383 as the benchmark, but I know full well I will never have a post bench of anything over 240 (it's impossible) or 120fps because of my monitor's refresh rate. Now along comes Miguel with the exact same system but with a VGA monitor that cannot hit even 45 fps. His benchmark is 383 also, but he can never exceed 45fps. So, in effect the benchmark lied to Miguel, because he thought he was doing as good as my system, but he can never get there.

Also, you can get two different results across DVI (-I or -D, or whatever) than you do HDMI.

[MrRiplEy[H]]

Well, that is until you realize that there can be quite a difference between internal benchmark and what is actually ever seen on screen. That number can tell you something along the lines of how everything pre-RAMDAC (or today it might be TMDS) turns out, but there is still a lot more beyond that depending on the monitor(s) you are using. For instance, I get 383 as the benchmark, but I know full well I will never have a post bench of anything over 240 (it's impossible) or 120fps because of my monitor's refresh rate. Now along comes Miguel with the exact same system but with a VGA monitor that cannot hit even 45 fps. His benchmark is 383 also, but he can never exceed 45fps. So, in effect the benchmark lied to Miguel, because he thought he was doing as good as my system, but he can never get there.

Also, you can get two different results across DVI (-I or -D, or whatever) than you do HDMI.

No. There are practical limits to current screen technology and its limitations have nothing to do with benchmarking the GPU performance. It's like saying you can't test drive a Porsche at the track because you only can drive 90mph on the highway.

[gyrene81]

i really hate to say it but...Ripley is right. looks like Chalenge may not have the full information on the function of the ramdac chip and how it affects gpu performance.