[H] Yet again say SLI is smoother than crossfire

[RussianSensation]

Keep in mind they are using a geometric mean for FPS.

I am fully aware of that and I read the comments section. Where did HD7950's 30-40% advantage disappear to in Skyrim? How did HD7950 boost go from being on par with GTX670 in frames per second in MOH to losing to a GTX660Ti with MSAA?

Here is a fun fact - a HardOCP review is used to show HD7970 CFX providing a worse experience without Radeon Pro and TechReport is used to show GTX660Ti providing a smoother experience over HD7950 950mhz. Seems legit, right? What does HardOCP say about HD7950 boost vs. GTX660Ti in their November review then?

"Middle pricing band – This pricing band was far less competitive as the Radeon HD 7950 with Boost simply demolished the GTX 660 Ti across the board with regards to raw frame rates and overall game play experience across our suite of testing." ~ Source

Am I seeing this correctly? HardOCP's review of HD7970CFX is used to demonstrate how it's inferior to GTX680 SLI but then HardOCP's review where HD7950 and I quote "demolished the GTX660Ti" is mysteriously omitted from the same thread. If the same people who are using HardOCP as their website to go to read all their other reviews, wouldn't this disagreement with TR's review strike them as contradictory and wouldn't they have noted this in the thread? Seems objective...

All I am saying is, let's at least look at the big picture. Is HardOCP's review of SLI vs. CFX an outlier from what we heard about SLI vs. CFX from countless other reviews online? No, it's not. TR's review is though. In this thread, both of these reviews are mixed together to prove a point about smoothness but we shouldn't at all mix them together. One of them, [H] is actually in-line with what we've known for years about SLI > CFX, at least without Radeon Pro. The TR review has results which are unusual to put it mildly, with sporadic performance and frames per second in at some games that make no sense whatsoever at 2560x1440 compared to countless other professional reviews.

 

[Vesku]

Yes, I really can't take the Techreport review seriously atm. Waiting to see how thorough their follow up is. Going to be disappointed if they don't do an update next week after dropping a hefty conclusion like this: " Instead, we have a crystal clear recommendation of the GeForce GTX 660 Ti over the Radeon HD 7950 for this winter's crop of blockbuster games."

 

[RussianSensation]

I still can't comprehend TR's Skyrim performance. It was almost a given for the last 10 months that if you wanted a card for BF3, you got GTX680 and if you wanted to play Skyrim, nothing touches the HD7970. HD7900's family advantage at 2560x1440 or 2560x1600 with AA is not just going to disappear in Skyrim.

This is a mid-Nov review, with Catalyst 12.11.

 

[SirPauly]

I wanted it compared to 2x680 not 7970CF as it doesn't cast any light on that hardware metering. I want to see if that hardware metering isn't some kind of marketing trick and that it is indeed some actual hardware and that it really has advantage over 2x680 in smoothness. Right now I can't see any special silicon on 690's PCB that could do that.

I think the hardware is in Kepler.

http://techreport.com/review/22890/nvidia-geforce-gtx-690-graphics-card/10


Has GTX 680 Sli, HD 7970, CrossFire, GTX 690.

 

[NIGELG]

Like I said before if only two review sites are drawing these conclusions then one has to wonder what is their motivation .....or is it that they made a mistake or are incompetent....Which is it??

 

[Keysplayr]

Like I said before if only two review sites are drawing these conclusions then one has to wonder what is their motivation .....or is it that they made a mistake or are incompetent....Which is it??

"or is it that they made a mistake or are incompetent or right on the money....Which is it??"

fixed a little. This is all we want to see. Fully aware that it could prove completely false or ring true.
And soon it will be 3 review sites. I'm sure Ryan Smith here at AT will conduct the experiment when he figures out the best way to approach this and decide the best tools he'll need.

 

[NIGELG]

"or is it that they made a mistake or are incompetent or right on the money....Which is it??"

fixed a little. This is all we want to see. Fully aware that it could prove completely false or ring true.
And soon it will be 3 review sites. I'm sure Ryan Smith here at AT will conduct the experiment when he figures out the best way to approach this and decide the best tools he'll need.

On whose money?Nvidia's money?....lol.

Why don't you respond to RS's detailed posts about the strangeness/credibility of the Techreport findings?Tear his analysis to shreds if you are so correct.

You are not doing this for the good of the community.You are grasping at straws because your precious Nvidia is lower in many benchmarks on many sites.You are an Nvidia focus group member through and through and could never be seen as objective or having 'noble' motivations.

Investigations might be done by more sites if they deem the topic worthy of such.Not because we listen to and heed your 'write your letter to your editor' campaign.

What review websites choose to investigate is fine by ME as long as it's scientific and objective as possible.

 

[Lepton87]

I think the hardware is in Kepler.

http://techreport.com/review/22890/nvidia-geforce-gtx-690-graphics-card/10


Has GTX 680 Sli, HD 7970, CrossFire, GTX 690.

If it is in kepler then 690 is not any smoother than 2x680 or 2x670 overclocked to the same performance level is that correct?

 

[raghu78]

On whose money?Nvidia's money?....lol.

Why don't you respond to RS's detailed posts about the strangeness/credibility of the Techreport findings?Tear his analysis to shreds if you are so correct.

You are not doing this for the good of the community.You are grasping at straws because your precious Nvidia is lower in many benchmarks on many sites.You are an Nvidia focus group member through and through and could never be seen as objective or having 'noble' motivations.

keys is not here to objectively argue or respond to RS posts. Do you think a Nvidia Focus Group member will say anything negative about Nvidia [H]'s review shows HD 7970 Ghz CF is allowing the games to be playable at higher settings than GTX 680 SLI. BF3 Multiplayer is playable with 2x MSAA on HD 7970 Ghz CF at 5760 x 1080. With GTX 680 SLI it only runs FXAA. BF3 singleplayer is playable with 4X MSAA on HD 7970 Ghz . With GTX 680 SLI it runs 2x MSAA. Sleeping Dogs is playable with High AA at 5760 X 1080 on HD 7970 Ghz CF. With GTX 680 SLI its runs Normal AA.

the smoother solution can't provide the better image quality. then what the heck is the smoothness for. also how much of this smoothness affects playability. does it mean the AMD solutions stutter such that you can't enjoy the playing experience. Surely that can't be the case. [H] would mention that. If the GTX 680 SLI provided the same image quality and was perceivably smoother then it has the edge. But now thats not the case.

 

[-Slacker-]

Looking at those graphs, I honestly don't get what they mean when they say the SLI 680s are smoother - I see more consistent frame rates in one or two graphs, but "consistent" isn't equivalent to "smooth" when the nvidia setup delivers maximum frame rates that are often below AMD's minimums;

If one setup oscillates between -say - 80fps maximums and 40fps minimums, then that's still smoother than the setup with fps ranging from 45fps to 30fps, even if the second system has fewer dips.

I am disappointed in the way [H] chose to conclude the article, but at least the raw data is still there for anyone to see.

 

[Hitman928]

Yes, I really can't take the Techreport review seriously atm. Waiting to see how thorough their follow up is. Going to be disappointed if they don't do an update next week after dropping a hefty conclusion like this: " Instead, we have a crystal clear recommendation of the GeForce GTX 660 Ti over the Radeon HD 7950 for this winter's crop of blockbuster games."

There have been several instances in Techreport reviews, especially conclusions, that have left me scratching my head which is why I was skeptical right away of their latest comparison. I give them credit for trying to do things differently and approach everything from a different angle, but sometimes I wonder how much they really want to be taken seriously. Have they responded in their forums/comments to people pointing out the obvious?

 

[Hitman928]

Looking at those graphs, I honestly don't get what they mean when they say the SLI 680s are smoother - I see more consistent frame rates in one or two graphs, but "consistent" isn't equivalent to "smooth" when the nvidia setup delivers maximum frame rates that are often below AMD's minimums;

If one setup oscillates between -say - 80fps maximums and 40fps minimums, then that's still smoother than the setup with fps ranging from 45fps to 30fps, even if the second system has fewer dips.

I am disappointed in the way [H] chose to conclude the article, but at least the raw data is still there for anyone to see.

Right now they have nothing more than a subjective take on it. They've mentioned that what they feel is happening doesn't show up in the numbers, but that the SLI rig just feels smoother, which is why there's another whole thread in this forum about how to actually make this stuff testable and repeatable. At least though, [H] has the sense to say that they can't quite figure it out yet and will continue investigating the matter.

 

[Keysplayr]

On whose money?Nvidia's money?....lol.

Why don't you respond to RS's detailed posts about the strangeness/credibility of the Techreport findings?Tear his analysis to shreds if you are so correct.

You are not doing this for the good of the community.You are grasping at straws because your precious Nvidia is lower in many benchmarks on many sites.You are an Nvidia focus group member through and through and could never be seen as objective or having 'noble' motivations.

Investigations might be done by more sites if they deem the topic worthy of such.Not because we listen to and heed your 'write your letter to your editor' campaign.

What review websites choose to investigate is fine by ME as long as it's scientific and objective as possible.

If you or anybody else find the TR report or H findings "strange" Then you all should be completely for other review sites to try and disprove them. No?
My goodness. Dig yourself deeper much? hehe.

 

[NIGELG]

If you or anybody else find the TR report or H findings "strange" Then you all should be completely for other review sites to try and disprove them. No?
My goodness. Dig yourself deeper much? hehe.

Read the last paragraph of my post that you quoted.

 

[SirPauly]

If it is in kepler then 690 is not any smoother than 2x680 or 2x670 overclocked to the same performance level is that correct?

Based on TechReport findings, the GTX 680 SLi and GTX 690 were indistinguishable but, personally don't know and why I asked nVidia.

 

[SirPauly]

Why don't you respond to RS's detailed posts about the strangeness/credibility of the Techreport findings?

Possibly, different areas and using Windows 8.

 

[Vesku]

There have been several instances in Techreport reviews, especially conclusions, that have left me scratching my head which is why I was skeptical right away of their latest comparison. I give them credit for trying to do things differently and approach everything from a different angle, but sometimes I wonder how much they really want to be taken seriously. Have they responded in their forums/comments to people pointing out the obvious?

They received quite a bit of requests for further testing in the comments to the article. Scott Wasson, Techreport Editor-in-Chief, says they are doing more tests.

 

[SirPauly]

the smoother solution can't provide the better image quality. then what the heck is the smoothness for. .

http://www.hardocp.com/article/2012/11/12/fall_2012_gpu_driver_comparison_roundup/5

Just offering an example based on you offered can't.

 

[Keysplayr]

http://www.hardocp.com/article/2012/11/12/fall_2012_gpu_driver_comparison_roundup/5

Just offering an example based on you offered can't.

"The GTX 680 was able to perform at 2560x1600 using the High Anti-Aliasing setting. While the HD 7970 appears to have the headroom in the frame rate to support that setting, the game play experience was not as fluid as we expected it to be. The GTX 670, HD 7950 and GTX 660 Ti were all able to perform at 1920x1080 with High Anti-Aliasing settings configured."

This^

It's kind of what I'd want to see put to the test by as many other review sites as possible. To either concur with H's findings, find a reason for this happening, or disprove it by some repeatable means. Whatever.

 

[Leadbox]

"The GTX 680 was able to perform at 2560x1600 using the High Anti-Aliasing setting. While the HD 7970 appears to have the headroom in the frame rate to support that setting, the game play experience was not as fluid as we expected it to be. The GTX 670, HD 7950 and GTX 660 Ti were all able to perform at 1920x1080 with High Anti-Aliasing settings configured."

This^

It's kind of what I'd want to see put to the test by as many other review sites as possible. To either concur with H's findings, find a reason for this happening, or disprove it by some repeatable means. Whatever.

Reading that in isolation, one would almost get the impression it was applied across the board, so I checked and it only applies to ONE game-Sleeping Dogs. These little oversights can weaken arguments, just saying

 

[Keysplayr]

Reading that in isolation, one would almost get the impression it was applied across the board, so I checked and it only applies to ONE game-Sleeping Dogs. These little oversights can weaken arguments, just saying

Nuh uh. From the Article in the OP, this was said about Battlefield 3

"We turned on 2X MSAA on both GTX 680 SLI and 7970 GHz Edition CrossFireX. AMD Radeon HD 7970 GHz Edition CrossFireX is 34% faster. However, also notice how much the framerate jumps all over the place with CrossFireX, whereas with GTX 680 SLI it is more consistent. This consistency helps provide GTX 680 SLI gaming with smoothness, whereas with 7970 GHz Edition CrossFireX we notice these wild fluctuations in framerate."

The oversight was you not seeing that it wasn't just in one game. But, that is what we are all here for, to help each other out.

 

[wand3r3r]

I am fully aware of that and I read the comments section. Where did HD7950's 30-40% advantage disappear to in Skyrim? How did HD7950 boost go from being on par with GTX670 in frames per second in MOH to losing to a GTX660Ti with MSAA?

Here is a fun fact - a HardOCP review is used to show HD7970 CFX providing a worse experience without Radeon Pro and TechReport is used to show GTX660Ti providing a smoother experience over HD7950 950mhz. Seems legit, right? What does HardOCP say about HD7950 boost vs. GTX660Ti in their November review then?

"Middle pricing band – This pricing band was far less competitive as the Radeon HD 7950 with Boost simply demolished the GTX 660 Ti across the board with regards to raw frame rates and overall game play experience across our suite of testing." ~ Source

Am I seeing this correctly? HardOCP's review of HD7970CFX is used to demonstrate how it's inferior to GTX680 SLI but then HardOCP's review where HD7950 and I quote "demolished the GTX660Ti" is mysteriously omitted from the same thread. If the same people who are using HardOCP as their website to go to read all their other reviews, wouldn't this disagreement with TR's review strike them as contradictory and wouldn't they have noted this in the thread? Seems objective...

All I am saying is, let's at least look at the big picture. Is HardOCP's review of SLI vs. CFX an outlier from what we heard about SLI vs. CFX from countless other reviews online? No, it's not. TR's review is though. In this thread, both of these reviews are mixed together to prove a point about smoothness but we shouldn't at all mix them together. One of them, [H] is actually in-line with what we've known for years about SLI > CFX, at least without Radeon Pro. The TR review has results which are unusual to put it mildly, with sporadic performance and frames per second in at some games that make no sense whatsoever at 2560x1440 compared to countless other professional reviews.

I love how the NV fanboys conveniently ignore this and pretend like they aren't making a mountain out of a molehill.

I say bring on the testing but some fanboys are just trying to create some distraction from the fact NV has lost the performance crown. NV isn't the fastest (single gpu) card in the town atm.

Quick pull up the blinders and ignore these in the supposed quest for objective information.

 

[Leadbox]

Nuh uh. From the Article in the OP, this was said about Battlefield 3

"We turned on 2X MSAA on both GTX 680 SLI and 7970 GHz Edition CrossFireX. AMD Radeon HD 7970 GHz Edition CrossFireX is 34% faster. However, also notice how much the framerate jumps all over the place with CrossFireX, whereas with GTX 680 SLI it is more consistent. This consistency helps provide GTX 680 SLI gaming with smoothness, whereas with 7970 GHz Edition CrossFireX we notice these wild fluctuations in framerate."

The oversight was you not seeing that it wasn't just in one game. But, that is what we are all here for, to help each other out.

Thats dual gpus you are on about now,

The sleeping dogs one you quoted was single gpu

Good thing thing that I can spot these oversights, eh?

 

[Lepton87]

Based on TechReport findings, the GTX 680 SLi and GTX 690 were indistinguishable but, personally don't know and why I asked nVidia.

I guess they didn't bother to answer that question but i'll ask anyway.
What was their answer?

 

[SirPauly]

I guess they didn't bother to answer that question but i'll ask anyway.
What was their answer?

As an advocate for smoother gaming, noticed the hardware frame metering wording for the GTX 690 and had a desire to know more about it. In the live pc perspective review and chat - I tried to garner some data about it from nVidia and basically received: " SirPauly, good to see you here, sorry, we're not here today to talk about hardware frame metering!"

In other words, received nothing but at least I did get a response.