Techreport 7950 vs. GTX 660 Ti "Smoothness" videos

[RussianSensation]

The graphs don't say anything about duration. In Hitman the 660 Ti renders fewer fps during the benchmark, hence the graph is shorter. The x-axis is not time, it is number of fps and thus dependent on performance. Lower performance -> shorter graph..

And thanks for pointing that out. I just realized I read the graph wrong. :thumbsup:

 

[VulgarDisplay]

My main issue is that they are using two of the least smooth game engines available to test smoothness. Gamebryo and Dunia are two of the buggiest engines on earth.

 

[HurleyBird]

It could just be power-tune acting up on that particular card, or it could be recent driver unintentionally (or intentionally, to boost FPS) broke frame metering. Would be nice to test either different cards, or the same methodology across several driver versions to see exactly where the issue materialized.

 

[boxleitnerb]

AMD responds:

AMD spokesman Antal Tungler is on the ball and said that the review had "raised some alarms" internally at the company. AMD is investigating and hoped to have some answers for us "before the holiday." It appears that AMD also expects the 7950 to perform well in FPS-based benchmarks and give the GeForce GTX 660 Ti a good kicking too.

http://www.fudzilla.com/home/item/29813-radeon-hd-7950-has-no-problems-with-windows-8

 

[Whitestar127]

My main issue is that they are using two of the least smooth game engines available to test smoothness. Gamebryo and Dunia are two of the buggiest engines on earth.

Strictly speaking it's called Creation engine, not Gamebryo. Although it's probably an updated Gamebryo engine.

It would be nice to see the two videos next to each other in normal speed also, just to see how the actual gaming experience is. In the end that is what matters.

 

[Keysplayr]

Strictly speaking it's called Creation engine, not Gamebryo. Although it's probably an updated Gamebryo engine.

It would be nice to see the two videos next to each other in normal speed also, just to see how the actual gaming experience is. In the end that is what matters.

You might not be able to pick up on the smoothness on a 29fps you tube video, but looking at it slowed down, you can surely see what contributes to gameplay with a more choppier feeling.

 

[Whitestar127]

I can see it there yes, but to me it's more interesting to see how it is in real-time.
Maybe film it with a 60 or 50 fps camera? I don't know.

Edit: It would also be nice to see comparisons to SLI, since I'm considering 670 SLI. Maybe there are such benchmarks available already?

 

[Keysplayr]

I can see it there yes, but to me it's more interesting to see how it is in real-time.
Maybe film it with a 60 or 50 fps camera? I don't know.

I've read they've filmed it with 120 and 240fps camera. Problem with You tube is, if there is a video taken at 240fps, and youtube only displays 29fps and slower, you would probably not have an accurate real time representation.
The TR videos however, offer much slower than 29fps playback so it is actually easy to see. Also easy to understand, after seeing the lag spikes, how somebody can say that it doesn't feel as smooth even though framerates are high.

 

[Silverforce11]

Playing Skyrim on 7850 and 7950 through multiple drivers, texture mods etc.. never noticed this kind of behavior. For TR to do a crazy bench that contradicts every other site and their own prior review is too sus. 660ti beating 7950 by huge % in games where every review has it being defeated by huge %? Come on, something is messed up with their rig.

 

[ICDP]

Playing Skyrim on 7850 and 7950 through multiple drivers, texture mods etc.. never noticed this kind of behavior. For TR to do a crazy bench that contradicts every other site and their own prior review is too sus. 660ti beating 7950 by huge % in games where every review has it being defeated by huge %? Come on, something is messed up with their rig.

I have been playing Skyrim using a 7970 and GTX 680, both of them locked at 60 FPS with Vsync and triple buffering. They do suffer some stutter and FPS drops on ocassion but nothing like the videos from TR. To be honest both cards in those videos show the exact symptoms that occur when games are run without vsync or triple buffering. Neither the GTX 660Ti or the 7950 are showing acceptable performance IMHO but the 7950 is much worse.

If there is a driver problem, and it is very possible given the information TR have posted, it will be sorted. It certainly isn't hardware, or ALL HD 7xx0 reviews would mention the problem. I just hope AMD don't take months to fix it if the drivers are the problem. They are betas so let's hope it's easily and promptly fixed.

 

[Silverforce11]

This is exactly my point, its a widely played game and on all sorts of hardware and in almost all reviews. For TR to have such opposite results in fps compared to everyone else suggests a) something with their hardware config or b) the particular scene they picked is a weakness for GCN on recent drivers.

As was the waterfall scenes in Crysis that was a weakness on 58xx and 69xx vs Fermi, as benched by AT here. If its b) I can believe it since it has occurred in the past in some scenes in games. It's something for the driver team to fix and is a legit issue.

ps. Seeing those videos in detail, its a stutter mess on both cards. The screen tearing is also blatant and intolerable.

 

[SirPauly]

AMD responds:


http://www.fudzilla.com/home/item/29813-radeon-hd-7950-has-no-problems-with-windows-8

Isn't that wonderful? The awareness gets noticed and has the attention of AMD!

 

[SirPauly]

While I admit that these results are a bit strange, you should keep in mind that different scenarios can yield different results. Unless you use the exact same scene, benchmarks will not necessarily be comparable. Maybe techreport picked a worst case scenario for AMD by chance, who knows?

Do we even know what scenes techspot and hardocp are benchmarking? Indoor, outdoor?

Indeed!

 

[Vesku]

Just throwing this out there, but is it possible that Nvidia is overriding some in game settings and syncing/buffering anyways? Curious if there is a way to confirm a card is obeying all in game settings.

 

[Keysplayr]

Playing Skyrim on 7850 and 7950 through multiple drivers, texture mods etc.. never noticed this kind of behavior. For TR to do a crazy bench that contradicts every other site and their own prior review is too sus. 660ti beating 7950 by huge % in games where every review has it being defeated by huge %? Come on, something is messed up with their rig.

What do you mean contradicts? No other site besides H is testing for anything like this as of yet. They are the first to address this and I'm sure this opens the door to this becoming standardized testing. Especially after AMD acknowledges it. Or "alarms" go off.

 

[ICDP]

Just throwing this out there, but is it possible that Nvidia is overriding some in game settings and syncing/buffering anyways? Curious if there is a way to confirm a card is obeying all in game settings.

Not likely to be honest. The 660Ti is also showing horrendous tearing and stuttering, it's just that the 7950 is even worse.

 

[flopper]

My main issue is that they are using two of the least smooth game engines available to test smoothness. Gamebryo and Dunia are two of the buggiest engines on earth.

if a gameengine done for consol and lazy programmers its a world of hurt for all.
drivers 12.7 showed 7900 seriers better than nvidia in smoothness and such.
might be a patch, or driver change or something else that alter the stuff.

game I play things just rock with a 7970.

 

[Vesku]

I don't competitive FPS so I play with VSync which I don't think would be effected by this at all. At least I haven't noticed anything and I'm a bit sensitive to multi-gpu stutter.

Edit: Also I can't replicate their Skyrim results but Sleeping Dog does have similar spikes at max settings. But, imo, Sleeping Dogs wouldn't be playable on either card for me without Vsync and settings that could hit 60fps consistently.

 

[The Alias]

What do you mean contradicts? Not other site besides H is testing for anything like this as of yet. They are the first to address this and I'm sure this opens the door to this becoming standardized testing. Especially after AMD acknowledges it. Or "alarms" go off.

but every site acknowledges that the 7950 delivers WAAYYYY more frames per second than the 660ti therefore it is contradictory

 

[Silverforce11]

What do you mean contradicts? Not other site besides H is testing for anything like this as of yet. They are the first to address this and I'm sure this opens the door to this becoming standardized testing. Especially after AMD acknowledges it. Or "alarms" go off.

Russian already covered this. If you don't care to read his posts with the answers, i certainly dont want to reflog all that data.

 

[wand3r3r]

Russian already covered this. If you don't care to read his posts with the answers, i certainly dont want to reflog all that data.

It just goes to show, certain fanbois won't even read posts which contain well thought out answers and ideas which happen to question the lack of proof on the subjective topic they propose.

I'm all for the testing and therefore improvement of experience, but can't any of the NVidiots even try to counter the arguments that seriously question what's going on with this so far single site that has 'uncovered' this potential phenomena.

Everytime Russian lays down the logic hammer the fanbois just crawl back into their (presumably) moms basements. Then when the fanboi comes and says something stupid or subjective they pile on on focus on that vs actually debating with facts.

Bring on the investigations, it will likely only improve my experience as an end user, however it's just humorous to see the fanbois scrambling to defend or tarnish the other company.

I'm mainly referring to Russians post as he has compiled the clearest list of contradictory evidence, but it's clear more investigation should be made.

 

[tviceman]

Never settle for consistent frame times, amirite? JK.

12.11's sacrifices consistency for throughput, plain and simple. My question is, did AMD knowingly do it in hopes of mainstream video card reviews sticking to average fps? Hardocp and techreport are throwing a wrench into AMD's current performance advantage and it's clear that some don't want to believe what is being reported.

 

[Jaydip]

From TR
"AMD spokesman Antal Tungler told us that our article had "raised some alarms" internally at the company, and he said they hoped to have some answers for us "before the holiday.""

If AMD themselves acknowledges the issue I don't see why we are arguing about it.

 

[KingFatty]

I recall reading something about how, for the video, he selected an outdoor scene with lots of vegetation etc. that would put more stress on the video cards.

Could that choice lead to why his results deviate from other websites tests?

 

[Hitman928]

I recall reading something about how, for the video, he selected an outdoor scene with lots of vegetation etc. that would put more stress on the video cards.

Could that choice lead to why his results deviate from other websites tests?

He retested the old scene and found the same stuttering issues, there was no stutter in previous tests.

From TR
"AMD spokesman Antal Tungler told us that our article had "raised some alarms" internally at the company, and he said they hoped to have some answers for us "before the holiday.""

If AMD themselves acknowledges the issue I don't see why we are arguing about it.

I think people are taking this out of context. This isn't AMD admitting there are stuttering issues present, it's more of, AMD is aware of their results and will do their own testing to verify if there is an issue. The good news, if it is an issue with the new drivers, AMD now has an early jump on getting it sorted out. If they can't replicate the issue, well, we'll know soon enough.