HD5970 Microstuttering tests

[rge]

Milliseconds per frame, as the graphs have been showing is the correct measurement.

Measuring frames per second is an average measurement over a period, not an instantaneous one. This is pretty basic math.

I think we are just going to have to agree to disagree, as its beating a dead horse now, so my last post on the subject, so feel free to get last word in.

You use the units you can understand, or you prefer, or you deem are the only correct ones, or whatever your issue is. I will continue to use fps and mph and any other measurement I care to as either instantaneous or average, depending on whether I take an instantaneous or average measurement, and I will let the average persons reading comprehension determine its use. If that bothers you, and apparently it does since you are complaining repeatedly about it, either dont read my posts, or convert my posts in your head to units you understand or like or deem only correct, or make your own posts in units you prefer. But just because you have a bizarre issue with the units being used, doesnt mean the rest of the world can no longer use them.

And I would give that same argument with next officer that gives you a ticket. Sorry officer miles PER HOUR is not an instantaneous measurement, it is an average over an hour, go follow me for an hour then we will talk...you seem to be able to do it with a straight face.

But to be perfectly honest, you would never be able to convince me, regardless of what you post, that you really believe what you just posted. My opinion will always be, 1) you first responded without reading the thread or my post, that is clear from your initial post, 2) you didnt like the response you got, 3) you found something else to argue about, semantics.

[MadDias]

I think we are just going to have to agree to disagree, as its beating a dead horse now, so my last post on the subject, so feel free to get last word in.

You use the units you can understand, or you prefer, or you deem are the only correct ones, or whatever your issue is. I will continue to use fps and mph and any other measurement I care to as either instantaneous or average, depending on whether I take an instantaneous or average measurement, and I will let the average persons reading comprehension determine its use. If that bothers you, and apparently it does since you are complaining repeatedly about it, either dont read my posts, or convert my posts in your head to units you understand or like or deem only correct, or make your own posts in units you prefer. But just because you have a bizarre issue with the units being used, doesnt mean the rest of the world can no longer use them.

And I would give that same argument with next officer that gives you a ticket. Sorry officer miles PER HOUR is not an instantaneous measurement, it is an average over an hour, go follow me for an hour then we will talk...you seem to be able to do it with a straight face.

But to be perfectly honest, you would never be able to convince me, regardless of what you post, that you really believe what you just posted. My opinion will always be, 1) you first responded without reading the thread or my post, that is clear from your initial post, 2) you didnt like the response you got, 3) you found something else to argue about, semantics.

u cannot see the microstutter issue when ur comparing fps.
look at dimitriman´s post... he showed u what microstuttering really is.
if u meassure fps and think u can tell that there is microstuttering involved, u totally misunderstood the phenomenon. such as most people that use the term "microstutter".
u have to record the time that each frame needs to be rendered. that is what the guys, that brought the whole "MS Myth" to life, did.
they compared two systems that pulled absolutely the same fps in a game. but the multi gpu system seemed to be stuttering when they reached a very low fps region around 30fps. and this was at a time when everybody was using a crt monitor.

if the time between rendering frames fluctuates even on single gpu systems, there is absolutely no point why games should run worse on a multi gpu system. and the whole microstutter bubble busts.

[rge]

u cannot see the microstutter issue when ur comparing fps.
look at dimitriman´s post... he showed u what microstuttering really is.
if u meassure fps and think u can tell that there is microstuttering involved, u totally misunderstood the phenomenon. such as most people that use the term "microstutter".
u have to record the time that each frame needs to be rendered. that is what the guys, that brought the whole "MS Myth" to life, did.
they compared two systems that pulled absolutely the same fps in a game. but the multi gpu system seemed to be stuttering when they reached a very low fps region around 30fps. and this was at a time when everybody was using a crt monitor.

if the time between rendering frames fluctuates even on single gpu systems, there is absolutely no point why games should run worse on a multi gpu system. and the whole microstutter bubble busts.

And yet another one that posts without reading previous posts.

I posted exactly what you just said in regarding to varying times to display frames how many times. I also graphed the varying framerates PER FRAME how many times? Go back and read the last page 9 of my posts, starting with 220, but then all on that page

[STEvil]

I think we are just going to have to agree to disagree, as its beating a dead horse now, so my last post on the subject, so feel free to get last word in.

You use the units you can understand, or you prefer, or you deem are the only correct ones, or whatever your issue is. I will continue to use fps and mph and any other measurement I care to as either instantaneous or average, depending on whether I take an instantaneous or average measurement, and I will let the average persons reading comprehension determine its use. If that bothers you, and apparently it does since you are complaining repeatedly about it, either dont read my posts, or convert my posts in your head to units you understand or like or deem only correct, or make your own posts in units you prefer. But just because you have a bizarre issue with the units being used, doesnt mean the rest of the world can no longer use them.

And I would give that same argument with next officer that gives you a ticket. Sorry officer miles PER HOUR is not an instantaneous measurement, it is an average over an hour, go follow me for an hour then we will talk...you seem to be able to do it with a straight face.

But to be perfectly honest, you would never be able to convince me, regardless of what you post, that you really believe what you just posted. My opinion will always be, 1) you first responded without reading the thread or my post, that is clear from your initial post, 2) you didnt like the response you got, 3) you found something else to argue about, semantics.

My truck gets 12 puddings per cat.

[blindbox]

Don't you guys know what instantaneous speed is? ds/dt. Now that's basic maths for you.

I'm not going to bother repost what rge has said. He got everything right.


Tearing occurs when the monitor caught the graphic card switching frames red handed. If you see tearing going from top to bottom, that means the monitor is jumping from one frameswitch to another frameswitch, although at a minor offset. Vsync tells the monitor to show a different frame only after the frame has finished rendering.. causing the input lag. Hence yes, you can experience tearing at fpses lower than your monitor's refresh rate (I'm speaking from experience, not theories).

I just wanted to answer a question eRazorzEDGE brought regarding tearing at fps < Hz .

So dual card wielders, any difference between a single card and a dual card on a high avg fps game? The thread so far has told me that some people can see it, some people don't.

[eRazorzEDGE]

I haven't been able to test for the last couple days due to multiple items going back for RMA (all started failing at the same time, horrible), but I'll try some out next week.

The screen tearing of the palm trees in Crysis is plain ridiculous once you can point a finger at it. It doesn't go away and then come back again, or only happen a few times in a minute. It almost appears like the screen is catching the GPU right in the middle of switching to a new frame EVERY time the screen refreshes, only I can only really tell with the palm trees and no other objects in game so far.

On higher FPS games, like L4D, I'll see the usual (or at least what I think is the usual) type of screen tearing where it's almost like a horizontal line across the screen where the image above and below that line don't match up. But that only happens every so often, not on a regular repeatable basis, at least to me. But we're way off topic...

... So it's all about high-speed film capture and some tests... who's got the $12k camera?

[gifaro]

I have a powerful gaming rig. I passed from a single GTX 285 to a SLI and then a TriSli. Each step gave me better graphic in games.Imo MS is an issue only for people who enjoy studying it.

[biohead]

guys, fps is just the derivative (dframes)/(dt) just like velocity is dx/dt and this measurement is instantaneous, since dt is infinitesimally small. when you're talking about average fps, however, dt is a selected period of time which is useless in measuring microstuttering.

it doesn't matter if you measure milliseconds (y) over frames (x) or fps (y) over time (x), they're all connected. and so fps or millisecond fluctuations on a microscale indicate microstuttering.

though i think measuring milliseconds between frames (y) over time (x) is the best way to do, since it's the most obvious.

[prava]

But it was under 60 frames per second. On the single GPU it was only 30 FPS. It happens consistently and is very reproducible.

I'm just not really getting exactly how it does it. It's my understanding that the display refreshes 60 times a second w/o any "output" or any sort of timing signal. I could be wrong on that however. If I'm right, how does just switching on Vsync, and getting below the monitors refresh rate in FPS, make the screen tearing go away?

That would mean that the system or GPU needs to know the exact time to render a frame so that monitor doesn't refresh while the single contains half of one frame and half of the next.

I guess I'll just have to buy a high speed camera and do some more tests.

That's easy work. Vsync works on direct multiples of your screen's refresh rates. Which means that if your screen is a 60hz one and you get below 60fps the system will change to the next multiple one, which in this case is 30fps. That is why you never have tearing with vsync no matter how bad your fps is because the distribution of frames among time is always flat.

There's a reason why the "review sites" list FPS (max, min, average), it's be cause it's a whole hell of a lot easier to understand than these graphs we're all producing. We don't even have any factual information on why, even on single GPU setups, each frame takes a different amount of time to render... even if the alternating high's and low's stay at their same respective values over a period of time.

Sure we can add this type of testing method to the traditional FPS tests, but these will never outweigh the traditional FPS tests. For right now, people should remain objective and try not to draw any conclusions from these graphs.

The fact of the matter is that traditional reviews don't say crap about wether the gameplay was smooth or not as FPS (as the reviews state it) doesn't give us anything. What would give us an idea about the quality of gameplay would be the biggest and lowest rendering time and a comparison. A game in which the average rendering frame time its 20ms and the lowest and highest only diverge in 10ms would be a game highly playable whether a game with 10ms average and 30ms of dispersion would not (call it ms or fps as rge said, know I think I understand what he was trying to say although I feel like using frame-render-time it's more correct and useful).

[jmke]

each frame takes a different amount of time to render

scene complexity ?

[STEvil]

scene complexity ?

Scene, physics, engine, drivers...