Techreport 7950 vs. GTX 660 Ti "Smoothness" videos

[RussianSensation]

So the 660TI should display more graphical glitches because they are running at even higher FPS than the 7950?

I'm not sure why disabling v-sync for both would make a comparison invalid.

As I understood the post, MrK6 is not saying to disable V-sync, the opposite. He is saying, you shouldn't manually disable the engine's V-sync in the .ini file since it causes more problems. In fact both GTX660Ti and HD7950 start stuttering in Skyrim, just 1 happens to stutter more, but neither is smooth.

If you turn off Vsync in-game, don't expect the game to work properly (Silverforce11 post explains how issues are found in the game's forum discussions due to this setting) because when NV and AMD optimize their drivers, they do so for the default settings in the .ini, or how a gamer would be playing it in the first place. If you turn off the Vsync like that, you throw off how the driver works for that particular game and then your results are not necessarily accurate. You are now assuming the Bethesda game engine itself is well optimized to handle the frame intervals and it might not work well depending on the driver/vendor used. Of course you can manually force Vsync with 3rd party tools then if you don't like how native frame metering/Vsync works in Skyrim by default.

Read this thread to see how much it impacts the HD7900 series stutter in Skyrim:
http://forums.anandtech.com/showthread.php?t=2289641

 

[SirPauly]

But TechReport did this in the past where AMD was much smoother in Skyrim -- now based on this review -- they're dishonest?

 

[blastingcap]

is it me or is everbody very ornery lately?

Some people are always like that. There is a ***** on this forum who is well-known as being biased and spewing one-sentence snark and spreading FUD rather than nailing things down in concrete terms. Like instead of answering a question straight up, he'll ask a one-line sarcastic question back. Can't stand him and wish he were banned, but due to his forum status and visibility, that will not happen anytime soon.

 

[AnandThenMan]

But TechReport did this in the past where AMD was much smoother in Skyrim -- now based on this review -- they're dishonest?

Let me guess, TechReport should be taken seriously because they said positive things about Radeon products in the past. Maybe the market should decide, instead of vocal review sites.

 

[SirPauly]

Anand says it better than me:

https://twitter.com/anandshimpi/status/279440323208417282

 

[Keysplayr]

This video is incorrect and so is TechReport, although I'm not surprised in the dishonesty in reporting or propogating it. Skyrim has a vsync option by default (called iPresentInterval=0 in the SkyrimPrefs.ini file). The Creation engine really depends on this and some pretty funky glitches happen if you turn it off and get high FPS. TechReport clearly turned this off since their cards are getting over 60FPS with a 69FPS average on the 7950 and 74FPS average on the GTX 660Ti. So not only are people trying to prove a point with only a single game, but what they're actually using is TechReport breaking the game and then providing dishonestly presented data.

So it was off for BOTH cards, is what you're saying? Or are you saying that the video we see is a result of different settings between the two cards? All benchmarks that show off fps have Vsync turned off anyway. AMD PR claims the Never Settle drivers dramatically increase performance across the board. That can't be tested with Vsync on, else everyone would get 59-60 fps. So, if you have to have Vsync enabled on the 7950, for example, and capped at 60 to get smoother gameplay but don't have to do this on the 660Ti, for example, then that negates the performance lead AMD PR claimed with never settle. Who cares when you need to cap framerates.

 

[VulgarDisplay]

So it was off for BOTH cards, is what you're saying? Or are you saying that the video we see is a result of different settings between the two cards? All benchmarks that show off fps have Vsync turned off anyway. AMD PR claims the Never Settle drivers dramatically increase performance across the board. That can't be tested with Vsync on, else everyone would get 59-60 fps. So, if you have to have Vsync enabled on the 7950, for example, and capped at 60 to get smoother gameplay but don't have to do this on the 660Ti, for example, then that negates the performance lead AMD PR claimed with never settle. Who cares when you need to cap framerates.

What they're saying is that if you turn off Vsync on Creation/Gamebryo the entire engine goes haywire. You'll see movable objects rocket around a room and other oddities. The game is not meant to run without Vsync on. It doesn't matter what brand you are using.

 

[Keysplayr]

What they're saying is that if you turn off Vsync on Creation/Gamebryo the entire engine goes haywire. You'll see movable objects rocket around a room and other oddities. The game is not meant to run without Vsync on. It doesn't matter what brand you are using.

Then explain why the 660Ti was still smoother under the same circumstances that the 7950 was subjected to. I don't see how "introducing" this new data has any relevance on the smoothness of the respective cards performance. Both cards had to endure the same conditions and one was much smoother than the other.

 

[Silverforce11]

Then explain why the 660Ti was still smoother under the same circumstances that the 7950 was subjected to. I don't see how "introducing" this new data has any relevance on the smoothness of the respective cards performance. Both cards had to endure the same conditions and one was much smoother than the other.

AMD doesn't optimize their drivers for Skyrim under situations where the game engine isn't functioning as its meant to? Hence "raising alarms" and looking into it.

 

[Keysplayr]

AMD doesn't optimize their drivers for Skyrim under situations where the game engine isn't functioning as its meant to? Hence "raising alarms" and looking into it.

Nvidia does then. Apparently. Seems they didn't suffer from the "malfunctioning game engine".

 

[Silverforce11]

Nvidia does then. Apparently. Seems they didn't suffer from the "malfunctioning game engine".

Not as much. Still pretty messed up in those vids.

 

[lavaheadache]

Coming from somebody who has been into this hobby for the better part of 2 decades, hardly do I ever rely on my out of the box experience with a given product to be the deciding factor. That isn't to say I don't value the importance of a great out of box experience.

Remember when the Far Cry came out and everybody was pushing the geforce 6 to be superior for that game because of HDR capability yet in order to get that feature you had to use console commands to enable it? Hardly something that a noob would no about.

Throughout the years there have been countless scenario's like that.

With that being said.....

-Run Afterburner or something similar

-set frame limit to 60 or 120for 120hz

-enjoy smooth gameplay (works a majority of the time). Side effect is that you don't use up power that would otherwise go to undetectable frames.

I *have* to do this with both single and multi gpu setups from both vendors.

/thread


And yes I did this aswell with my 1 month long ownership of an Evga GTX 680.

 

[Siberian]

is it me or is everbody very ornery lately?

Some people are upset at being proven wrong. This video evidence ruined the credibility of several posters.

 

[lavaheadache]

Nvidia does then. Apparently. Seems they didn't suffer from the "malfunctioning game engine".

Key's I found Skyrim to be jittery on my 580 and 680 just as I do with my 7970. I have alway's felt the need to run it with a frame cap.

As to the degree of jitteriness I can really compare directly as I never had them running side by side on equal screens.

 

[Rikard]

So you must break the game by not supported tweaks to see the stuttering? That explains why I never observed this. I tried a number of different games and benches and I cannot see any stuttering anywhere with either nvidia or AMD. Sounds like the typical storm in a cup of water we have on these forums...

 

[Ferzerp]

So you must break the game by not supported tweaks to see the stuttering? That explains why I never observed this. I tried a number of different games and benches and I cannot see any stuttering anywhere with either nvidia or AMD. Sounds like the typical storm in a cup of water we have on these forums...

It amuses me when people are ultra skeptical about a rather large site's findings, but are willing to accept the word of john q forumposter on any speculation that they did something incorrect based on no evidence, and just a few claims.

 

[Keysplayr]

It amuses me when people are ultra skeptical about a rather large site's findings, but are willing to accept the word of john q forumposter on any speculation that they did something incorrect based on no evidence, and just a few claims.

And so it must be.

 

[Elfear]

And so it must be.

There can only be one.

 

[Keysplayr]

There can only be one.

Why do I always picture Shatner saying this?

 

[Elfear]

Why do I always picture Shatner saying this?

Good question. Christopher Lambert might be offended you mistook him for Shatner.

 

[Blastman]

It amuses me when people are ultra skeptical about a rather large site's findings, but are willing to accept the word of john q forumposter on any speculation that they did something incorrect based on no evidence, and just a few claims with no evidence.

The problem lies with TR's flawed methodology. Did you look at the BF3 results from TR I posted earlier in this thread? The results from just that 1 game are all over the place. If TR just posts one graph from BF3 using the results from Rock and a hard place what should we conclude from that? AMD has much smoother game play in BF3? AMD has smoother drivers in games in general over NV? It's nonsense -- they wouldn't be valid conclusions from the limited data given the wide variances that their methodology is susceptible to.

And then there is this little tidbit from THG …

THG …

We've run a number of other benchmarks and need to add, for the sake of fairness, that Nvidia's advantage is only evident if its driver is optimized for smooth frame rates rather than raw performance. The company takes certain apps, like synthetic benchmarks and commonly-tested games, and tweaks them to yield higher numbers at the expense of consistency.

So, how many games has NV tweaked their drivers for "smooth" frame rates … 5, 6 …10 … a few more? Is TR going to disclose what games they benchmarked that NV has tweaked their drivers for smoother frame rates? Nope, don't see it mentioned on TR. Is TR going to tell us further that outside a small set of games that NV has tweaked their drivers for, that for the other 99% of games out there, there is generally no difference between NV and AMD for smooth frame rates? Nope, don't see any disclosure about that there either.

All these fancy ms traces and graphs and talk of stutter are virtually meaningless if your methodology is completely flawed, you don't have enough results to make valid conclusions, and you just happen to leave out some pertinent facts.

 

[Keysplayr]

<------ confused. Wouldn't it be a very "good" thing when a company "tweaks" drivers for smoother gameplay? Isn't that a thing somebody would want? Just curious.

 

[Zanovar]

Good question. Christopher Lambert might be offended you mistook him for Shatner.

hahaha

 

[KingFatty]

How can we know which games a video card has been optimized for? Does that show up in the drivers?

It would be pretty cool if the driver had a setting for frame metering and you could turn it off or on, and customize that setting on a per-game basis (like have it be part of the profile settings).

Why in the world haven't video card makers jumped all over this as an option?

 

[thilanliyan]

The problem lies with TR's flawed methodology. Did you look at the BF3 results from TR I posted earlier in this thread? The results from just that 1 game are all over the place. If TR just posts one graph from BF3 using the results from Rock and a hard place what should we conclude from that? AMD has much smoother game play in BF3? AMD has smoother drivers in games in general over NV? It's nonsense -- they wouldn't be valid conclusions from the limited data given the wide variances that their methodology is susceptible to.

And then there is this little tidbit from THG …



So, how many games has NV tweaked their drivers for "smooth" frame rates … 5, 6 …10 … a few more? Is TR going to disclose what games they benchmarked that NV has tweaked their drivers for smoother frame rates? Nope, don't see it mentioned on TR. Is TR going to tell us further that outside a small set of games that NV has tweaked their drivers for, that for the other 99% of games out there, there is generally no difference between NV and AMD for smooth frame rates? Nope, don't see any disclosure about that there either.

All these fancy ms traces and graphs and talk of stutter are virtually meaningless if your methodology is completely flawed, you don't have enough results to make valid conclusions, and you just happen to leave out some pertinent facts.

QFT...unfortunately, this MS problem seems to differ from game to game which makes investigating the issue that much harder. Hopefully more sites can/will take a crack at it and see if there really is an issue there.