Techreport 7950 vs. GTX 660 Ti "Smoothness" videos

[KingFatty]

Has TR done any more testing of any other games besides Skyrim? Sounds like they got the setup and everything ready to roll, but rolling has not yet commenced...?

 

[thilanliyan]

Equally, the determination to make it seem like a general AMD = stutterfest while Nvidia = smooth is getting tiresome. TR got some contradictory results compared to their own previous testing, so it might be drivers or CAPS or a bad card but it most certainly isn't proof that ALL AMD HD 7xx0 cards = stutterfest.

QFT. Skepticism =/= denial. Until there is more proof from several other reputable websites, I'm afraid rumours, FUD, etc will be perpetuated depending on who is perpetuating.

 

[3DVagabond]

I'm not saying it's a mountain, I'm saying it raises questions and we should have more games tested, and with various settings and resolutions, to see if it's a real problem or an isolated occurrence. Playback of high speed footage is to let you see more clearly why a game felt less smooth, and to correlate it with frame time spikes. I think we can all agree that the more testing done the better--it's foolish to rely one one game at one setting and resolution. Ideally we'd have tests of every game at every conceivable setting and resolution.

I'm not trying to say that the measured results are wrong. There is definitely more frames showing higher latency with the 7950 than the 660ti in their tests. Slowed down to 3fps (judging with the naked eye) makes it noticeable. If at 60fps (or 40fps or 30fps) you can't see it, what's it matter? Add that the settings used were completely unplayable, with the screen tearing, and it's even less compelling.

 

[Keysplayr]

Exactly. But that analysis needs to be done before straight away concluding that Nvidia has won 'hands down' as is the attitude of a number of posters here, also others who state it is impossible that AMD video cards may have a situational lack in smoothness. I'm all for rigorous analysis, but until that analysis comes we have a very specific situation that people are either trying to outright deny or extrapolate to some team-based win, and that is irrationality and hyperbole at its best.

And you actually care that somebody says that. Can't you just be satisfied with your own opinion and let others have theirs?

 

[Aristotelian]

And you actually care that somebody says that. Can't you just be satisfied with your own opinion and let others have theirs?

No, because posters like me focus on truths of this world, not seek to distract casual viewers of this forum with sensational comments that have no basis in fact or the actual world that we live in. I'm not the only one, in case you haven't noticed, who promotes rigorous fact finding methodologies. Why don't you take your own advice, and perhaps benefit from learning the difference between 'respecting someone's opinion' and 'separating fact from fiction' - I think you'll find that the latter option is far more productive.

 

[Rikard]

If this problem is so wide spread, why aren't these guys throwing up all kinds of videos of it in a variety of games? Instead we get a video of Skyrim. I naturally hitchy, buggy game. I game which I love never the less.

I'm all for exposing shoddy practices and what not so I would like to have more proof. So where are the videos?

Well, I think it is every 7950 and 660ti owner's responsibility to test the claims of TR and report back what they see. I still cannot see anything but butter smooth performance in Skyrim which is quite different from what you and TR are seeing. Since I cannot replicate their preconditions, nor their results, I suspect something is wrong with what they did.

However I could reproduce this,

although my frame times are about 2/3 of what they have, translating into 50% more FPS. But the spikes are there in that game (measured in Fraps), so that is no lie! Still I do not see anything. I must be immune to stuttering.

I think that concludes my participation on this topic: God knows what they did to make Skyrim perform so poorly, I can confirm spikes in at least one other game, but for me personally it is irrelevant.

 

[Keysplayr]

No, because posters like me focus on truths of this world, not seek to distract casual viewers of this forum with sensational comments that have no basis in fact or the actual world that we live in. I'm not the only one, in case you haven't noticed, who promotes rigorous fact finding methodologies. Why don't you take your own advice, and perhaps benefit from learning the difference between 'respecting someone's opinion' and 'separating fact from fiction' - I think you'll find that the latter option is far more productive.

Are you on Crack or something? Truths of this world? Sensational comments that have no basis in fact? You my friend have just changed the dictionary definition of delusional. You believe TR's findings are all one giant untruth. A lie. A scam. Are you serious? You know what, don't answer that. I know it already. We will have a shot in your name tonight at the local pub.

 

[Homeles]

Equally, the determination to make it seem like a general AMD = stutterfest while Nvidia = smooth is getting tiresome. TR got some contradictory results compared to their own previous testing, so it might be drivers or CAPS or a bad card but it most certainly isn't proof that ALL AMD HD 7xx0 cards = stutterfest.

"Bad card" and "bad rig" have been ruled out. TTR has stated that they have achieved similar results when swapping with a 7950 and swapping for a new rig.

 

[ICDP]

"Bad card" and "bad rig" have been ruled out. TTR has stated that they have achieved similar results when swapping with a 7950 and swapping for a new rig.

Thanks for the info. Many people and other review sites get different results using their 7950s. [H] recommend the 7950 as the best card for price/perf out of a few of their recent reviews using these very same drivers. Many people have posted timeframe charts showing their HD 7xx0 cards do not have any more stutter than equivalent GTX cards. Yet there are some here who outright claim those people are making stuff up. I now agree that simply stating "I see no stutters = nothing to report" is not a good way to prove that HD 7xx0 do or do not have problems with microsturter. The flip side of this coin is stating TRs results as an absolute fact and assuming they could not possibly be wrong.

There is no doubt TR are getting stutters from their 7950s. The only pertinent question is why? Other review sites and many individuals are not seeing or reporting the same problem. If this problem is as clear cut as AMD = stutters then it would not be only TR highlighting the problem for single GPUs.

 

[thilanliyan]

Are you on Crack or something? Truths of this world? Sensational comments that have no basis in fact? You my friend have just changed the dictionary definition of delusional. You believe TR's findings are all one giant untruth. A lie. A scam. Are you serious? You know what, don't answer that. I know it already. We will have a shot in your name tonight at the local pub.

Keys you are taking some posters comments here WAY too seriously and generally to be somehow against this investigation. The TR review is the ONLY review we have...until other sites test and come to the same conclusion, what's wrong with being skeptical of it (not outright dismissing it)? You seem to interpret anyone questioning their results as them dismissing it completely, which is not the case.

You yourself said a little paranoia is okay if it isn't misplaced. Well, the same applies to skepticism from the other side...what is wrong with that?

 

[antihelten]

To Rikard and ICDP, seeing as you both tested Skyrim, and observed the same stutter as TR, I was curious if you could produce frame time graphs similar to the ones TR does, when they zoom in on a small section (50 frames) of the run:

http://techreport.com/r.x/radeon-win8/skyrim-7950-w7.gif

The reason I ask is to see whether you can observe the same very peculiar pattern TR gets, where they got a slow frame (either 21 or 26 ms frame time), an average frame (17 ms frame time) and then a fast frame (7 or 12 ms frame time), repeated over and over (only one frame, frame 915, falls out of this pattern).

 

[ArchAngel777]

Keys you are taking some posters comments here WAY too seriously and generally to be somehow against this investigation. The TR review is the ONLY review we have...until other sites test and come to the same conclusion, what's wrong with being skeptical of it (not outright dismissing it)? You seem to interpret anyone questioning their results as them dismissing it completely, which is not the case.

You yourself said a little paranoia is okay if it isn't misplaced. Well, the same applies to skepticism from the other side...what is wrong with that?

No kidding. We should wait for more results... What do you say Keys? Maybe we shouldn't just take bad news against the competition and run with it without question... Ahh, but I see who I am talking too, or rather, what, so I guess my expectations are too high to expect any reasonable discussion.

With that said, the initial results should alarm anyone with an ATI card. I read this review and could easily see that ATI was lagging in the videos. I have a 7970, so this matters to me. But since I am not particularly vested in AMD or nVidia I have no problem with more investigation on the issue. If true, it is dissapointing. But my guess is that it may be a driver related issue, or possibly something with a specific configuration. I hope I am right.

The good news is that so far I have not noticed anything like those videos on my own rig. Buttery smooth.

 

[ICDP]

To Rikard and ICDP, seeing as you both tested Skyrim, and observed the same stutter as TR, I was curious if you could produce frame time graphs similar to the ones TR does, when they zoom in on a small section (50 frames) of the run:

http://techreport.com/r.x/radeon-win8/skyrim-7950-w7.gif

The reason I ask is to see whether you can observe the same very peculiar pattern TR gets, where they got a slow frame (either 21 or 26 ms frame time), an average frame (17 ms frame time) and then a fast frame (7 or 12 ms frame time), repeated over and over (only one frame, frame 915, falls out of this pattern).

I don't know how to create such a graph. I can upload the .csv files for you to check them out if you want. The .csv files are included from two tests in Skyrim, 1 with default CCC settings and the other with flip queue size set to 1 in RadeonPro. With flip queue at default the game is unplayable, setting it to 1 removed the stutters.

http://www.sharebeast.com/0enz23uan93n

 

[Rikard]

To Rikard and ICDP, seeing as you both tested Skyrim, and observed the same stutter as TR, I was curious if you could produce frame time graphs similar to the ones TR does, when they zoom in on a small section (50 frames) of the run:

http://techreport.com/r.x/radeon-win8/skyrim-7950-w7.gif

The reason I ask is to see whether you can observe the same very peculiar pattern TR gets, where they got a slow frame (either 21 or 26 ms frame time), an average frame (17 ms frame time) and then a fast frame (7 or 12 ms frame time), repeated over and over (only one frame, frame 915, falls out of this pattern).

Like this?

I have a different pattern, looks like a two stroke engine! Either the frames are the same speed, or a very fast frame is followed by a slower frame. The really interesting thing is that it looks like the fast frame is always coming first after a period of constant frame rate, rather than the other way round. The slow frames themselves are corresponding to about 40 FPS which probably why I do not perceive these as stuttering.

Edit: Actually at the 2 second mark there is a 3 frame periodicity for a while, while usually I see a 2 frame periodicity. That is more like what was shown in TR.

 

[ICDP]

Like this?

I have a different pattern, looks like a two stroke engine! Either the frames are the same speed, or a very fast frame is followed by a slower frame. The really interesting thing is that it looks like the fast frame is always coming first after a period of constant frame rate, rather than the other way round. The slow frames themselves are corresponding to about 40 FPS which probably why I do not perceive these as stuttering.

Interesting result. Can you try with flip queue size set to 1 using RadeonPro. I am interested to see if it makes such a massive difference as it did on my 7970 test run.

Thanks

 

[lavaheadache]

Like this?

I have a different pattern, looks like a two stroke engine! Either the frames are the same speed, or a very fast frame is followed by a slower frame. The really interesting thing is that it looks like the fast frame is always coming first after a period of constant frame rate, rather than the other way round. The slow frames themselves are corresponding to about 40 FPS which probably why I do not perceive these as stuttering.

that looks terrible.



How are you guys producing those graphs? I know how to get a list of the frame times out of fraps but what are you doing to make the graph?

 

[Rikard]

that looks terrible.



How are you guys producing those graphs? I know how to get a list of the frame times out of fraps but what are you doing to make the graph?

http://sourceforge.net/projects/frafsbenchview/
Every person have different sensitivity to stuttering. I suppose I am blessed with low sensitivity. Maybe my brain has vsync on its input from the retina.

 

[ICDP]

http://frapsforum.com/threads/new-tool-for-viewing-fraps-benchmark-files.2121/

 

[Rikard]

Interesting result. Can you try with flip queue size set to 1 using RadeonPro. I am interested to see if it makes such a massive difference as it did on my 7970 test run.

Thanks

I can do just for you! I tried before in Heaven but with my overclock that setting caused a driver failure, and even with factory OC only I got artifacts in Heaven, so it is nothing I consider a permanent solution... Just give me a sec.

 

[Rikard]

Wow... now, if I only could get that setting to work nicely in Heaven...

 

[ICDP]

Those match my findings exactly. I don't think it is a driver or hardware issue, if it was we wouldn't be able to fix Skyrim with some minor driver tweaks.

I think AMD need to pay more attention to individual game profiles and add more options such as flip queue size to the CCC as standard. The majority of people would not know to use 3rd party tools such as RadeonPro or NvidiaInspector to get the best from their games. At least Nvidia give far more user control out of the box with their standard CP options.

 

[3DVagabond]

Wow... now, if I only could get that setting to work nicely in Heaven...

So, if I understand correctly, this is just from queue size set to 1? No frame capping?

 

[Rikard]

So, if I understand correctly, this is just from queue size set to 1? No frame capping?

Well, the tweak to turn off Skyrims internal vsync does not work for me, possibly due to all the hacks and mods I use. So no, my FPS is a constant 60 no matter what I do.

For what it is worth, changing queue size also for Heaven gives somewhat more consistent frame rates, but not as nice as in Skyrim. And the frame periodicity is 6 in Heaven, 4 11 ms, then one frame at 17 ms followed by one frame at 7 ms.

 

[Homeles]

There is no doubt TR are getting stutters from their 7950s.

See, the important thing is that you've recognized this. Things over here at AnandTech tend to be a bit more mature than over at OCN, where I normally lurk. We had people (as in >1) over there claiming that Nvidia forced Scott Wasson's hand in this.

My theory is that Scott Wasson is having an affair with Jen-Hsun Huang.

The only pertinent question is why? Other review sites and many individuals are not seeing or reporting the same problem. If this problem is as clear cut as AMD = stutters then it would not be only TR highlighting the problem for single GPUs.

Exactly. And I would stick with phrasing that as "not reporting" the problem, because most sites don't comment on smoothness. I know SweClockers is looking into it and are going to publish a review, and they've already mentioned that they have found similar issues, although far less pronounced. Then there's the apparent investigation going on over at AMD that would be interesting to here the results of.

I think the most surprising thing about all of this is how big it's blown up. Regardless of the outcome of this, the resonations from this were substantial enough to get AMD's attention, spur fanboys into murderous frenzies, and earn myself some more infractions for pointing out how childish they were being, as I normally do.

 

[3DVagabond]

See, the important thing is that you've recognized this. Things over here at AnandTech tend to be a bit more mature than over at OCN, where I normally lurk. We had people (as in >1) over there claiming that Nvidia forced Scott Wasson's hand in this.

My theory is that Scott Wasson is having an affair with Jen-Hsun Huang.
Exactly. And I would stick with phrasing that as "not reporting" the problem, because most sites don't comment on smoothness. I know SweClockers is looking into it and are going to publish a review, and they've already mentioned that they have found similar issues, although far less pronounced. Then there's the apparent investigation going on over at AMD that would be interesting to here the results of.

I think the most surprising thing about all of this is how big it's blown up. Regardless of the outcome of this, the resonations from this were substantial enough to get AMD's attention, spur fanboys into murderous frenzies, and earn myself some more infractions for pointing out how childish they were being, as I normally do.

I just want to comment on this one point. It would be completely irresponsible for AMD to ignore this. Them looking into it shouldn't be misconstrued as endorsing anything.

To be clear, I'm not accusing you of this.