Spin Off: AT's Testing Methods & Uber Mode

[OILFIELDTRASH]

Anyone remember when there was a limited edition extreme oc gtx 460ee was tested against the new reference amd cards? I thought that was curious just as the thought of testing a card in quiet mode only is quite curious.

 

[TerryMathews]

Bravo...well said and I agree wholeheartedly.

This whole conversation could have been avoided by a little extra engineering.

Lol. Its AMDs fault that AnandTech is using inconsistent review procedures?

Please.

 

[Abwx]

Oh I don't know. I just got done reading a controversy at tomshardware that was picked up by probably 30 other web news outlets that is based directly on the terrible cooler on the 290 series. And then, I see noise mentioned as a negative in 100% of every web review of the 290 series. And then practically every PC enthusiast forum has pages upon pages of people complaining about the noise and how they wouldn't consider a 290 series card because of noise. Says who? Indeed. Was all of this entirely, 100% preventable? Yes, it could have been prevented - No throttling, no noise complaints, no user experience complaints, no conspiracy theories about reviews at AT.. Preventable. Yes, yes it was. But wasn't.

But, let's forget all that. AMD is a victim here, right? I mean AMD didn't need to prevent all of this with a better design.

Measurements protocols must stick to reality wich is not always
the case for some reviews.

As an exemple THG place their microphone at 0.5 metre , about
1.5 ft, from the card obviously in open environment wich is never
the case (pun intended) in real life situation.

On the other hand Hardware.fr place the mike at 0.6 metre
from their reference casing that is closed , far more accurate
way of doing estimations that stick to real life conditions.

The results ? Well , all those noisy useless debates about
alleged deafening levels are torned apart , i will add that
if you double the distance from 0.6 to 1.2 metres the sound
pressure is reduced by 6dB and so on.

http://www.hardware.fr/articles/910-10/bruit-temperatures.html

 

[Saylick]

I skimmed through the thread and didn't find anyone mentioned how the 6990 had an AUSUM switch, or that the 5970 was "guaranteed" to run at 5870 clock speeds even though it was advertised to run at a lower frequency.

In both cases, AT reviewed the cards and included OC versions for each of their benchmarks even though the cards weren't shipped using with AUSUM mode ON and the clock speeds raised to 5870 levels, respectively. That is a big difference than how they left the 290/290X in Quiet mode, even though if they would have followed the same protocol as the 5970 and 6990, they would have included Quiet and Uber mode in the test.

 

[ViRGE]

http://forums.anandtech.com/showthread.php?p=35694414&highlight=#post35694414

"What I see is double standards everywhere. Forum folks, reviewers.

Lets see what Ryan Smith said at the GTX 580 launch (44.4db Idle - 57.1db Load).

Load
Now lets see what he just said about the R9 290 (39.5db Idle - 57.2db Load).

Load
Consistency, that is."

Looking at the articles, it looks like they switched cases, CPUs/coolers, and benchmarks between then and now. The old noise data was only taken from FurMark, whereas they're now primarily focusing on a gaming workload (Crysis 3).

They don't have a 580 in their current lineup, so you can't compare a 580 by their modern test methodology to the 290. But they do have a 480, and the 480 was quieter than the 290 with Crysis 3. I'd assume the 580 would be quieter yet.

Anyone remember when there was a limited edition extreme oc gtx 460ee was tested against the new reference amd cards? I thought that was curious just as the thought of testing a card in quiet mode only is quite curious.

The blowback for that was quite severe. I'm sure that's part of the reason they only include cards at their default settings.

http://www.anandtech.com/show/3988/the-use-of-evgas-geforce-gtx-460-ftw-in-last-nights-review

 

[Vesku]

They don't have a 580 in their current lineup, so you can't compare a 580 by their modern test methodology to the 290. But they do have a 480, and the 480 was quieter than the 290 with Crysis 3. I'd assume the 580 would be quieter yet.

So 3 decibels results in a rant in the conclusion? Again the issue is not in citing the noise it's the way it was cited.

 

[ViRGE]

So 3 decibels results in a rant in the conclusion? Again the issue is not in citing the noise it's the way it was cited.

Well 3dB versus a 480; like I said against a 580 it would certainly be more. My point being that you can't compare this to old articles because the equipment and methodologies have changed. So you have to compare it to what everything gets under the current methods. It's 4db louder than 290X, and louder still than 780 and Titan. Which isn't to say that it's a bad card (or that I agree with AT), but that's quite a bit of a difference.

 

[Janooo]

Anyone remember when there was a limited edition extreme oc gtx 460ee was tested against the new reference amd cards? I thought that was curious just as the thought of testing a card in quiet mode only is quite curious.

There is nothing new here.
There was NV bias then and there is NV bias now.
All is well
Just remember to filter is out when you read AT review.

 

[blackened23]

There is nothing new here.
There was NV bias then and there is NV bias now.
All is well
Just remember to filter is out when you read AT review.

I don't get the "AMD can do no wrong" and victim mentality I see around here.

Noise is only part of the equation. The bigger part of the equation is that the 290 cards SEVERELY THROTTLE when using any sort of quiet fan profile. But, apparently, AMD can do no wrong. Whatever. You guys can argue over noise semantics and that stupid freaking BIOS switch all day long but it doesn't change the fact that to maximize performance you need a noisy card, and if you want a quiet card well you're screwed. Unless you want to severely impact performance. That BIOS switch should not even exist. AMD should have put a more versatile shroud as to give uber level performance with quiet mode noise, and that would obviously eliminate the need for controversy and the BIOS switch altogether - I don't expect half the people in this thread to acknowledge that because some people think AMD can do no wrong.

You know, I hope that the controversy blows up even further - apparently THAT is the only way AMD can get it the message that they shouldn't compromise on user experience with bone headed design decisions. I say this as someone who was genuinely excited for the 290 launch and wanted it to be AMD's "home run" GPU. Performance wise they hit the mark but as far as user experience goes? I cannot ignore the throttling and discrepancies between quiet and uber mode - A cooler that causes throttling by 250+ mhz at quiet fan profiles, well, yeah, that's a bone headed design decision that 100% COULD have been prevented - I can't believe AMD didn't want to make a better shroud with that in mind. Yes, I know some people will use water and some people don't care. A far great number of folks just want a GPU that they can "set it and forget it". Maybe an internet controversy is the only way for AMD to wisen up, and then maybe they can release "B" revision cards without crap blowers.

 

[Teizo]

Default vs default. You have to enable Uber Mode. It is not the default mode for the 290X. So, you use straight out of the box experience vs straight out of the box experience.

Then you can also bench in Uber Mode and then note the noise tradeoff, etc.

But, if AMD wanted reviewers to only test in Uber Mode for benchmarks...they should have stated so.

If you compare the default experience between both cards, the Ti is leaps and bounds better than the 290X. To compare Uber Mode of the 290X vs the default mode of the 780Ti would be unfair because basically Uber Mode is an overclocked mode. In other words, whatever % Uber Mode allows the 290X to run...then that is what you would have to add in addition to the 780Ti.

Pretty simple. Not Ryan's fault AMD had to raise the temp and fan limit on Hawaii to be able to compete with Kepler. And, not Ryan's fault AMD had to create a 'quiet' mode to keep people's hearing safe from possible hearing loss. It was AMD's decision to put a turd cooler on their card.

And by the way, whoever in AMD's marketing department that approved the term 'Uber Mode' needs a new job. Dumbest, gimmickiest name I've ever seen for a card just about.

 

[Erenhardt]

@Teizo,
SLI doesn't work out of the box. You need to do hardware modification to make it work. On the other hand Crossfire works out of the box.
If nvidia dared to ask for special test procedure in AMD 6870/6850 review. It is hard to believe they didn't do this again.

 

[Gikaseixas]

Reviewers just need to start testing in real world scenarios to better give the reader the closest to reality impressions. Open Bench testing was implemented for ease of use since they just have to switch cards, cpu's, ram, etc. I can see all the trouble they would have to go through by testing in a conventional case.
Problem with that is, open benching won't give you accurate noise readings and in some instances blower coolers do perform better inside a case.
Just saying...

 

[GaiaHunter]

Oh I don't know. I just got done reading a controversy at tomshardware that was picked up by probably 30 other web news outlets that is based directly on the terrible cooler on the 290 series. And then, I see noise mentioned as a negative in 100% of every web review of the 290 series. And then practically every PC enthusiast forum has pages upon pages of people complaining about the noise and how they wouldn't consider a 290 series card because of noise. Says who? Indeed. Was all of this entirely, 100% preventable? Yes, it could have been prevented - No throttling, no noise complaints, no user experience complaints, no conspiracy theories about reviews at AT.. Preventable. Yes, yes it was. But wasn't.

But, let's forget all that. AMD is a victim here, right? I mean AMD didn't need to prevent all of this with a better design.

And that is why the stores are in such a frenzy offering massive discounts to try to move the 290X/290...

The GTX 480 and the GTX 470 were much much worse values (well the 470 got a substantial discount after a while becoming less expensive than a 5870) than the 290X and 290, and still sold.

The GtX 580 and 570 were worse than the 290/290X are vs their respective competition and were massive successes.

People that actually bought the 290X/290 seem pretty satisfied with their cards.

So, no, the enthusiast segment didn't change that much. the enthusiasts get 400W bios for their cards.

 

[EliteRetard]

I'd also like to point out that 60db is normal conversational level noise...

 

[VirtualLarry]

If nvidia dared to ask for special test procedure in AMD 6870/6850 review. It is hard to believe they didn't do this again.

This.

 

[Will Robinson]

Teizo:

And, not Ryan's fault AMD had to create a 'quiet' mode to keep people's hearing safe from possible hearing loss.

 

[Carfax83]

@Teizo,
SLI doesn't work out of the box. You need to do hardware modification to make it work. On the other hand Crossfire works out of the box.
If nvidia dared to ask for special test procedure in AMD 6870/6850 review. It is hard to believe they didn't do this again.

That's a really terrible analogy. An SLI bridge is critical for the function of SLI. It cannot work properly without it.

A higher performance BIOS on the other hand merely supplements and enhances the card's capability, which is really no different than any overclocking software, except that it's at a deeper level and from the manufacturer.

Anandtech was right to test both the default and uber mode..

 

[TerryMathews]

Anandtech was right to test both the default and uber mode..

Exactly. What you seem to be missing is Ryan's statement that this was an exception and in the future only quiet mode would be tested and used for comparison purposes.

The issue isn't the review, it's the stated policy.

 

[Imouto]

Looking at the articles, it looks like they switched cases, CPUs/coolers, and benchmarks between then and now. The old noise data was only taken from FurMark, whereas they're now primarily focusing on a gaming workload (Crysis 3).

They don't have a 580 in their current lineup, so you can't compare a 580 by their modern test methodology to the 290. But they do have a 480, and the 480 was quieter than the 290 with Crysis 3. I'd assume the 580 would be quieter yet.

And that's the point. The test was done before and now running Furmark but Ryan tried to show the R9 290 in the worst possible light including the GTX 480 in the recent R9 290 review shooting himself in the foot badly, the scores then (GTX 580 review) and now (R9 290 review):

GTX 480 = 64.1db / 65db
HD 5870 = 59.3db / 61.3db
GTX 580 = 57.1db / X (not loud)
R9 290 = --- / 58.5db (unreasonable noise)

I hope you can solve the X.

 

[Janooo]

I guess here is the reason why Uber mode was not considered. 290X is apparently faster. These are the AT's numbers.
http://forum.beyond3d.com/showpost.php?p=1803944&postcount=1993

 

[railven]

Seems I'm not AT's audience. Guess wanting a thorough review of a product and the data to carry over is not what AT's aim is.

I got no problem reading "this thing is loud" but there is an issue with "I'm throwing half the data I got away because it requires a switch to be flipped."

I wonder how a HOF review would have been done here. "This card is no more special than any other custom cooled GTX 780."

 

[AlNasty]

And that is why the stores are in such a frenzy offering massive discounts to try to move the 290X/290....

Please link the stores, I can't seem to find any less than suggested retail, some asking for more than retail. Sold out in many situations.

 

[chimaxi83]

And, not Ryan's fault AMD had to create a 'quiet' mode to keep people's hearing safe from possible hearing loss.

When people post foolish crap like this, it's hard to take anything from them seriously.

 

[railven]

The blowback for that was quite severe. I'm sure that's part of the reason they only include cards at their default settings.

http://www.anandtech.com/show/3988/the-use-of-evgas-geforce-gtx-460-ftw-in-last-nights-review

Interesting, things that stand out to me is them stating Nvidia is more aggressive than AMD in getting review criteria. Also, it was AT's call on including the OC'ed 460, of course they did it on their own...this following after stating Nvidia is more aggressive?

I like where he said more information is better, but when I read the 780 Ti article, it sort of ommitted information that was in the article itself...

Still not saying anything about bias or whatever, I still love AT and will continue reading here (even if being told to go else where, since it seems I'm not the audience (go figure)) but this article's conclusion struck me as odd when I was digesting the charts.

If you'd rather not draw a conclusion from the Uber mode scores, don't throw them into the chart, but you'll still catch flak so I guess it's a lose/lose situation.

 

[raghu78]

When people post foolish crap like this, it's hard to take anything from them seriously.

exactly. I believe AT went overboard with this noise complaint. the R9 290 cards are not loud by any stretch of imagination until 55% fan speed.

http://www.youtube.com/watch?v=QQhqOKKAq7o

55% fan speed at 0:19 - 0:27. thats open test bench. in a closed case it would be even lesser.

hardocp and techreport have already said they found 50 - 55% fan speed not loud and reasonable to use for long durations.

http://techreport.com/review/25602/amd-radeon-r9-290-graphics-card-reviewed/9

"Much has been made of the R9 290X's relatively high power draw, operating temperatures, and noise levels. Obviously, the R9 290 shares these same characteristics, with a somewhat louder default fan profile. In my view, the only one of these properties that's really worth fussing over is the noise, since it's the thing you'll notice in day-to-day use.


We're apparently going to have to face this price/performance-versus-acoustics tradeoff for a while, so I spent some quality time with the R9 290 trying to get a handle on what I think of the noise, beyond the readings on the decibel meter. I've gotta say, there are some mitigating factors. For one, I like AMD's choice to stick with a blower that exhausts hot air out of the case rather than going for a triple-fan cooler that doesn't. I've seen those fan-based aftermarket coolers perform poorly in multi-GPU configs, and they often occupy quite a bit more space—maybe even a third expansion slot—in order to work their magic. I'm also not convinced AMD's cooler is a poor performer and therefore noisy, as some folks seem to think. Remember, it has more heat to remove than any of the coolers on the other cards we tested. Finally, I don't think this blower's ~49 dBA reading is the worst of its type. The quality of the sound isn't grating. Subjectively speaking, there are much more annoying coolers in this territory on the decibel meter. The impressively smooth, gradual ramp of fan speeds up and down in the new PowerTune algorithm helps make the noise less noticeable, too. This ain't an FX-5800 Ultra, folks."