AMD's smoothness factor put to the test by AMD & HardOCP...

[Cold Fussion]

Why aren't there any plots showing the rendering time for each frame on the two different systems? Saying one system feels smoother than the other is meaningless without any data being presented.

[Final8ty]

Why aren't there any plots showing the rendering time for each frame on the two different systems? Saying one system feels smoother than the other is meaningless without any data being presented.

Because its about how humans feel and perceive things and not about absolute numbers and on top of that it would not be a blind test.

[Blaber]

Why aren't there any plots showing the rendering time for each frame on the two different systems? Saying one system feels smoother than the other is meaningless without any data being presented.

We are talking about smoothness , do you have any app that can measure smoothness , because A system may play game at 90Fps with stuttering but B system may play game at 59fps but with no stutter , I dont know how you can make this out with fps meter e.t.c

[Cold Fussion]

We are talking about smoothness , do you have any app that can measure smoothness , because A system may play game at 90Fps with stuttering but B system may play game at 59fps but with no stutter , I dont know how you can make this out with fps meter e.t.c

Everything you perceive is directly from what is being displayed on the screen. The only difference to the end user is what is being display on the screen, so it's important to map this to correlate the data to the subjective impression.

[STEvil]

We are talking about smoothness , do you have any app that can measure smoothness , because A system may play game at 90Fps with stuttering but B system may play game at 59fps but with no stutter , I dont know how you can make this out with fps meter e.t.c

See here: http://www.xtremesystems.org/forums/...ity-Lost/page4

EDIT

And if anyone wants to be let in on a secret: There are a multitude of things that cause microstutter which can be perceived as a smoothness issue.

[Blaber]

Yes, it's called minimum FPS. Stuttering means disruption of the flow, that is, min FPS drops to below 25-30FPS, the level which the human brain can perceive.

A 2500k has been proven link to produce consistently better min FPS and average FPS than a 8150. A 2700k would walk all over it.

If anything, this shows how subjective human perception is, especially in a controlled environment by a interested party.

See answer below

And if anyone wants to be let in on a secret: There are a multitude of things that cause microstutter which can be perceived as a smoothness issue.

^ This Min fps , change in Fps are just few of the factor affecting smoothness , SB being able to put better min fps can only help that much.

[savantu]

We are talking about smoothness , do you have any app that can measure smoothness , because A system may play game at 90Fps with stuttering but B system may play game at 59fps but with no stutter , I dont know how you can make this out with fps meter e.t.c

Yes, it's called minimum FPS. Stuttering means disruption of the flow, that is, min FPS drops to below 25-30FPS, the level which the human brain can perceive.

A 2500k has been proven link to produce consistently better min FPS and average FPS than a 8150. A 2700k would walk all over it.

If anything, this shows how subjective human perception is, especially in a controlled environment by a interested party.

[Cold Fussion]

Yes, it's called minimum FPS. Stuttering means disruption of the flow, that is, min FPS drops to below 25-30FPS, the level which the human brain can perceive.

A 2500k has been proven link to produce consistently better min FPS and average FPS than a 8150. A 2700k would walk all over it.

If anything, this shows how subjective human perception is, especially in a controlled environment by a interested party.

It's not just minimum fps. If the fps is all over the place between 25 and 60, it's not going to be a smooth experience.

[Final8ty]

Yes, it's called minimum FPS. Stuttering means disruption of the flow, that is, min FPS drops to below 25-30FPS, the level which the human brain can perceive.

A 2500k has been proven link to produce consistently better min FPS and average FPS than a 8150. A 2700k would walk all over it.

If anything, this shows how subjective human perception is, especially in a controlled environment by a interested party.

It seems that you don't know what micro stuttering is, which is the timings between frames on multi GPU systems and nothing to do with the minimums, of course frame rates of 30-25 are not going to seem smooth, buts that's not micro stutter.

I have had a version of EVE that i capped to a solid 120fps as running at 400fps is a waste, but it only felt like 30fps because of the micro stuttering, had the same in early drivers for COD4, BFBC2, BF3.
Its also been shown that having a slight CPU bottleneck can help smooth out frame times.

[Final8ty]

Apparently it was HardOCP who set up the PC's.

[keithlm]

Repeat this test with fps caped at 60 and ppl can't tell anything.

Since you don't like the results you want to configure the systems so you can get the answer you prefer?

I think other people were just discussing unethical, illogical, and other shady things that could be done to bias results. Your suggestion would fit
perfectly into that category.

Actually it is interesting how so many people who obviously don't like these results feel the need to jump in and try to justify why a blind test is somehow "not real". Often these are the same people that will not question the bias of a reviewer creating a brand new review and including results from a 6 year old game or some older version of an application that has 3 or 4 newer versions. (That are now optimized for newer architectures.)

[Hornet331]

Since you don't like the results you want to configure the systems so you can get the answer you prefer?

I think other people were just discussing unethical, illogical, and other shady things that could be done to bias results. Your suggestion would fit
perfectly into that category.

Actually it is interesting how so many people who obviously don't like these results feel the need to jump in and try to justify why a blind test is somehow "not real". Often these are the same people that will not question the bias of a reviewer creating a brand new review and including results from a 6 year old game or some older version of an application that has 3 or 4 newer versions. (That are now optimized for newer architectures.)

oh boy here we go again... and seeing who posted I am not surprised at all...

You know that this would help both systems right? Especial when we have such cases:
http://www.hardocp.com/image.html?im...hfNV81X2wuZ2lm

Guess which on of the two will be "smoother"... but yeah its so unethical, illogical and shady to actually want to suggest a level playing field... well I'll be damned. Guess its intels fault that there cpu can spit out more fps in none gpu limited scenes and gets hit harder if gpu limit kicks back in..

But if you think AMD doesnt suffer from this take a look at this (even if its is SLI)...

http://www.hardocp.com/article/2011/...mance_review/5

With GeForce GTX 580 SLI this game mimics what we saw in DiRT 3. Average performance is very close between the two CPUs. However, the big difference is an inconsistency in performance similar to what we experienced in ARMA 2. So while the performance is similar, the smoothness and consistency is certainly better with the Intel 2500K.

If the minimum FPS is well below 60, you could run into laggyville...I had this with BF3, Vsync on

Maxfps is not the same as vsync. With vsync on it always tries to run at 60fps, if it fails to it runs at 30fps stable and so on and often this makes it worse then vsync off. maxfps just sets an upper celling at a certain fps rate.
What you eliminate with this is the huge fps drops lets say going form 120fps to 40fps and then back up to 90fps. Instead you go 60-40-60.

Just look at the Hardocp pic I linked at imagine everything cut off above the green 60fps line.

[naokaji]

It seems that you don't know what micro stuttering is, which is the timings between frames on multi GPU systems and nothing to do with the minimums, of course frame rates of 30-25 are not going to seem smooth, buts that's not micro stutter.

The blind test AMD and HardOCP did involved no multi gpu system.
And while they used the integrated Gpu with the low end rigs they used the same Gpu for the "proper" Pc, so if there was micro stuttering involved it would have had to come from somewhere else than the Gpu.

[Final8ty]

The blind test AMD and HardOCP did involved no multi gpu system.
And while they used the integrated Gpu with the low end rigs they used the same Gpu for the "proper" Pc, so if there was micro stuttering involved it would have had to come from somewhere else than the Gpu.

If there is any stutter it will not be MicroStutter.