The Cause Of And Fix For Radeon R9 290X And 290 Inconsistency

[littleleo]

AMD should have used a better quality fan in there reference builds. They seem penny-wise and dollar foolish. They have the cost delta to equip the cards with a very good cooling solution and they didn't. It gives me pause to buying new AMD cards when they launch. It seems best to wait a month so the bugs get worked out and new drivers are released.

 

[Quaddro]

Stick to the topic dudes..

Fanboyism make all people here seems happy to show their own stupidity...😀

Nvidia or AMD will not pay you even if you kill each others..

#Grab popcorn

 

[DrKlahn]

So, you didn't notice that they have the data and clearly, nothing happened here...

Not sure how you are coming to that conclusion. I'm not stating there is no issue.

Honestly, I am a man who always wants the truth to be out and not being burried. I cannot force people to be the same. And if you believe that the truth doesn't represent you, then I am afraid you are at the wrong site.

Who said anything about burying the truth? What about being 100% upfront with your findings. Does the truth that this issue is not one if you run the card in it's performance mode not matter? Do you feel that a buyer of a highend card was being done a service with how this was handled?

]I beg your pardon, but I think AMD choose to use an "quiet" mode in a flagship card. And I agree with you that we don't need quit modes in high end gaming. So why AMD even created a quiet mode profile?

If your need shifts to acoustics the option is there. We're both in agreement that it's pretty frivolous. But reviewers always raise issues with it, so we get a Quiet mode and the confusion that comes with it.

]
Why do you think they handled it badly?

When you are bringing what could be a potentially damaging issue to light, I have always found that the proper course is to publish the issue along with the facts you have to back it up. I understand Tom's wanting to make their readers aware of it, but they stated the issue with essentially no hard data to back it up. So everyone reading that article was left with the impression that AMD was purposefully misleading the press and the retail cards had serious problems. The reality was quite different. The conservative mode was a bit more conservative based on card vendor. Again if the card was tested in Uber mode you would have had no "crisis".

 

[JackNaylorPE]

"don't think gamers lacking headphones will be perturbed by the sound"

at 2200 and 2650 rpm are you kidding me ????

IIRC, it was said in the original 290 article that the sound level was just under 60 dBA before this increase. A ancient vacuum produces just 70 dBA .... new ones much less ...who wants to be in a room with a vacuum (or 2 in CF) running

So w/ 2 of these in CF I can expect H100i like noise levels ? Here's the H100's 2 fans @ 2000 rpm (1:45 mark)

http://martinsliquidlab.org/2013/03/12/swiftech-h220-vs-corsair-h100i-noise-testing/

hits 2650 @ 2:10mark

I don't see water cooling as the be all and end all solution others seem to ..... if ya need to put 2600 rpm fans on the radiators too what do we accomplish ? The H100i produces the same noise levels at those 2200 and 2650 fan speeds.

Bad move AMD, never shuda released this card with these coolers; should waited for your partners to come up with something better. No one is going to buy these knowing that quieter solutions are weeks away and having web sites issue "No Buy" recommendations is not where ya wanna be in the pre-holiday season. Great concept, great way to squeeze performance out of the hardware, great price, terrible planning and implementation.

I never understand why ANYONE will buy the reference design cards.... even for the GTX780, the aftermarket cooler is way better and quieter than the stock fan.

Well, reference card usually has the highest component quality..
That means more durable and has longer life than common 3rd party cards (xfx, zotac, his, polor, and another lowly cheap brand)..
.

Actually the reverse is true. The aftermarket factory overclocked cards not only have better coolers but with few exceptions (EVGA SC series for example) they have beefed up PCBs and beefier VRMs making then far more durable than the "reference cards" with weaker VRMs.

 

[Darkresurrection]

I never understand why ANYONE will buy the reference design cards.... even for the GTX780, the aftermarket cooler is way better and quieter than the stock fan.

Because to a CPU overclocker, the aftermarket cooler is crap? I just had the same style cooler on another card cause a factory-built PC overheat its CPU. As in thermal throttling. And the CPU was liquid-cooled!

The type of cooler your praising blows heat into the case, rather than out of the case. They're just a bad design concept done right, as opposed to a good design concept that's poorly executed.

I think it depends on the cooler design and how your system build is. Even when I was Bitcoin/Litecoin mining 24/7 my CPU never went above 32c idle or 50c under regular load, which is what it runs around normally.

I have a HD7970 Vapor-X just for reference.

I never understand why ANYONE will buy the reference design cards.... even for the GTX780, the aftermarket cooler is way better and quieter than the stock fan.

Well, reference card usually has the highest component quality..
That means more durable and has longer life than common 3rd party cards (xfx, zotac, his, polor, and another lowly cheap brand)..

i've still running my 4 years old 5850 reference card..overclocked to 1ghz since out of box, and cooled by mk-13..well, no issue until today, and still rock..

So then getting a Sapphire would be equal to or better than getting a reference board since Sapphire designs the reference board.

And as for higher quality components, not always true. The Sapphire Dual-X uses the reference board (on the HD7970) but the Vapor-X uses a custom PCB that has their black diamond chokes which are considered way better than the reference setup. As well it has more VRAMs which means a more stable voltage etc.

The reference is a good starting point but companies like Asus and Sapphire, for example, normally make better than reference designs.

do you wanna start the war again really!? AMD is not a thief like Intel or Nvidia, AMD cares for enthusiasts more than Intel or Nvidia...

If you are talking about overpriced products: AMD FX 8-Core Black Edition FX-9590 $829.99 plus shipping¹

I am not surprised to see this article at Toms, they don't like AMD and who knows why, I have my r9-290 delivered today, didn't notice such a joke, Toms is being ridiculous...

Really? In case you didn't notice Tom's Hardware gave an award at 290X, while in a few days, it gave nothing to the faster, colder, quiter 780ti.
And you should be glad that people like Chris are doing this, so you can have your 20% performance back, cause AMD reacted with new drivers after Chris (and maybe other reviewers too) told them about the problem.

who do I thank!? Chris!? I thank AMD because it let me max out crysis 3, enable every single option in that demanding game and enjoy the super smooth experience only for 600$ fx 8350+r9 290= 600$ it's a shame gtx titan is 1000$ + an intel 240$ cpu=1200$ now who cares for enthusiasts? Time to grow up lads hadn't it been for AMD, you would buy your now 240$ cpu, for 1000$ and your 1000$ gpu, for 2000$ go thank AMD for that...I was testing my card no problem as such, maybe AMD sent me the press card huh!? Toms was that brazen to find that issue alone...Nobody knew about it, not AMD not Sapphire, Not Asus, Toms knew about it much to your surprise, that awar could have been a part of the plan, so that people like you can reply here...Disappointing... well I know I am going to be banned, you are good with AMD you are banned, no wonder everyone supports Toms here

No one cares if you prefer AMD, Intel or SiS etc. That's a personal opinion.

My question is why does everyone always pair Intel with NVidia when they do it like you have?

Honestly I haven't run NVidia in a long time, I do well with ATI/AMD. But I prefer Intels CPUs as they have better drivers, do more driver updates and normally perform a bit better while using less power.

The fact that you are getting all uppity over THG finding an issue and going to AMD is a sign that you are blind. Its a fact that in quiet mode some of the retail GPUs were having issues causing them to run at lower clocks. AMD admitted it and fixed it.

Still why anyone would go for a reference card instead of waiting for a superior after market cooling card is beyond me. Bragging rights?

who do I thank!? Chris!? I thank AMD because it let me max out crysis 3, enable every single option in that demanding game and enjoy the super smooth experience only for 600$ fx 8350+r9 290= 600$ it's a shame gtx titan is 1000$ + an intel 240$ cpu=1200$ now who cares for enthusiasts? Time to grow up lads hadn't it been for AMD, you would buy your now 240$ cpu, for 1000$ and your 1000$ gpu, for 2000$ go thank AMD for that...I was testing my card no problem as such, maybe AMD sent me the press card huh!? Toms was that brazen to find that issue alone...Nobody knew about it, not AMD not Sapphire, Not Asus, Toms knew about it much to your surprise, that awar could have been a part of the plan, so that people like you can reply here...Disappointing... well I know I am going to be banned, you are good with AMD you are banned, no wonder everyone supports Toms here

Since you are a fanboy, trying to prove you that an FX8350 + R9 290 cannot max out Crysis 3, is pointless.

Also I see that telling you to thank people like Chris that making companies (I don't care which one) to realise and accept their problems and try to fix it instead of ignore it, is pointless also.

But to imply that people at Tom's Hardware hate or have something against AMD, while you have no proof and while AMD admitted their problem and released a driver to fix it, I am sorry but you are a delusional fanboy.

nah...you are, watch this... http://www.youtube.com/watch?v=XE4H5-kbOls ...and ok I am a fanboy, just look at the other posts, they have not had this problem either so I am not alone, now you see what delusion means...and what do you mean!? speak English please, you mean I can't max out crysis 3 on fx 8350+ r9-290!? well, it is sad because I can, and I play it smoothly, you can be sad about that, cause you have spent more and you cant!? well it is really sad, but be honest with people

You posted a video with a R9-290X. A r9-290 is not the same, its a bit weaker.

And Crysis 3? Meh.

nah...you are watch this... http://www.youtube.com/watch?v=XE4H5-kbOls ...and ok I am a fanboy, just look at the other posts, they have not had this problem either so I am not alone, now you see what delusion means...and what do you mean!? speak English please, you mean I can't max out crysis 3 on fx 8350+ r9-290!?

Since you don't trust Tom's Hardware that gave an Elite award to 290X:
Yeah I barely called maxing out unless the ideal for you is not 60FPS but 40...
Also since you don't know what delusion/al means in English: Delusion

And now about the video. He tests 3dMark so I don't see the point of linking it. Also the mighty reviewer lets say he is more red minded than clear minded... :
91XtremeRT

heat means nothing. amd is more stable, hence why amd has the overclock world record. amd is the only cpu to reach 5.0 ghz out of the box. power consumption????? lol. the price difference between amd and intel, and with the increases electricity use on amd, it would take 30 YEARS of use for the amd to surpase the intel in price. do the math, its fact.. you dont have any valid or factual points in yur entire statement. more power consumption means nothing if that increase is 50 cents a month lol.

And I can type more but lets just stop this. It isn't going anywhere is it?
You are not taking solid proofs from respectable sites and you take a 1000views youtube video that the uploader has an AMD evolve background and talks like this in his comments as a trustworthy evidence...

Since when 3d mark is not an authentic benchmark!? red minded!? huh!!! you are angry, I can smell your confusion...is you gtx titan doing any better in the benchmark you sent!?

3DMark is and always has been considered synthetic and never a real mark of real world performance. I have seen times when NVidia beats AMD and in the real world, its the opposite and vice versa.

No real gamer considers 3DMark a good benchmark to show a GPUs real performance.

Stick to the topic dudes..

Fanboyism make all people here seems happy to show their own stupidity...😀

Nvidia or AMD will not pay you even if you kill each others..

#Grab popcorn

no, Jimmy listen...Tomshardware could have done this better, could have used better literature, not that picture and not that humiliating sentence,the literature you use to criticize is different from the one Tomshardware has obtained look Jimmy, Tomshardware could have chosen a title as such http://www.extremetech.com/gaming/170636-well-that-was-quick-amd-solves-r9-290-throttling-problem-with-a-new-driver ... not that childish picture and the naive sentence to show its hatred towards AMD or whatever, I might be blind, I have grown up with AMD, since I was 14 AMD is in my chassis, I have experienced topnotch moments with AMD, it reminds me of all those days and nights playing games, AMD vs Nvidia...AMD vs Intel...it's the same story of game consoles, no one will step down..but even the blind can tell prejudice and criticism apart... The way toms addressed the problem was not mature, like it or not...

 

[Novuake]

no, Jimmy listen...Tomshardware could have done this better, could have used better literature, not that picture and not that humiliating sentence,the literature you use to criticize is different from the one Tomshardware has obtained look Jimmy, Tomshardware could have chosen a title as such http://www.extremetech.com/gaming/170636-well-that-was-quick-amd-solves-r9-290-throttling-problem-with-a-new-driver ... not that childish picture and the naive sentence to show its hatred towards AMD or whatever, I might be blind, I have grown up with AMD, since I was 14 AMD is in my chassis, I have experienced topnotch moments with AMD, it reminds me of all those days and nights playing games, AMD vs Nvidia...AMD vs Intel...it's the same story of game consoles, no one will step down..but even the blind can tell prejudice and criticism apart... The way toms addressed the problem was not mature, like it or not...

Oh the IRONY!!!!

 

[youssef 2010]

A slight difference in fan speed can translate to a huge temperature difference. I remember that my 4870 used to hover around 80 C when using the default fan speeds. After the venerable Catalyst 9.2 was released, increasing the fan speed by only 6% caused the card to drop to mid 60s. It seems that the same weird feature still applies to AMD's cards five years later

 

[Darkresurrection]

what do you mean Novuak?

 

[Memnarchon]

Who said anything about burying the truth? What about being 100% upfront with your findings. Does the truth that this issue is not one if you run the card in it's performance mode not matter? Do you feel that a buyer of a highend card was being done a service with how this was handled

Since quiet mode is an official mode from AMD, what's wrong with that? As I said before I agree with you that people might play in uber mode but this is AMD's fault for doing it from the begining, not Tom's Hardware who they found something wrong about it.

If your need shifts to acoustics the option is there. We're both in agreement that it's pretty frivolous. But reviewers always raise issues with it, so we get a Quiet mode and the confusion that comes with it.

So now the option is there but we shouldn't make it a big deal, if it isn't working properly...

When you are bringing what could be a potentially damaging issue to light, I have always found that the proper course is to publish the issue along with the facts you have to back it up. I understand Tom's wanting to make their readers aware of it, but they stated the issue with essentially no hard data to back it up. So everyone reading that article was left with the impression that AMD was purposefully misleading the press and the retail cards had serious problems. The reality was quite different. The conservative mode was a bit more conservative based on card vendor. Again if the card was tested in Uber mode you would have had no "crisis".

Who said "purposefully"? Accidents do happen though...
Would that change the fact that press cards and retail cards dont work the same at quiet mode?

 

[Darkresurrection]

Who said anything about burying the truth? What about being 100% upfront with your findings. Does the truth that this issue is not one if you run the card in it's performance mode not matter? Do you feel that a buyer of a highend card was being done a service with how this was handled

Since quiet mode is an official mode from AMD, what's wrong with that? As I said before I agree with you that people might play in uber mode but this is AMD's fault for doing it from the begining, not Tom's Hardware who they found something wrong about it.

If your need shifts to acoustics the option is there. We're both in agreement that it's pretty frivolous. But reviewers always raise issues with it, so we get a Quiet mode and the confusion that comes with it.

So now the option is there but we shouldn't make it a big deal, if it isn't working properly...

When you are bringing what could be a potentially damaging issue to light, I have always found that the proper course is to publish the issue along with the facts you have to back it up. I understand Tom's wanting to make their readers aware of it, but they stated the issue with essentially no hard data to back it up. So everyone reading that article was left with the impression that AMD was purposefully misleading the press and the retail cards had serious problems. The reality was quite different. The conservative mode was a bit more conservative based on card vendor. Again if the card was tested in Uber mode you would have had no "crisis".

Who said "purposefully"? Accidents do happen though...
Would that change the fact that press cards and retail cards dont work the same at quiet mode?

The problem is solved, now they are the same at quite mode, read the article again, the variance of 3% had already been announced by AMD, and that wont chnage anything

 

[Darkresurrection]

I am just rotfl because you think th is paid off by nvidia. Sorry that th did not put powdered sugar and sprinkles on the review for you. the 290x may be a great value but most people do not want a jet engine in there pc, well no one i know.
http://www.youtube.com/watch?v=T5MbOGoEMDY

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

jet engine!!!? why can't i hear anything from my chassis then!? I gotta call AMD and thank them for sending their best model for me :lol: Jet engine....your chasses are made of paper gentlemen? my chassis is here http://www.green-case.com/products/case/midi/case.php?model=x-9 is it because it is silent ready that I cant hear even a whisper

 

[bigpinkdragon286]

While I appreciate the hoopla Tom's is making about the issue, if everyone is really so concerned with getting to the bottom of this, why can't I find any charts showing the chip running under 100% fan load? It's not about noise at this point, it's about seeing if the chip can hit and maintain it's targeted max frequency with consistency, and if it can, it's pretty clear the cooling is insufficient, at which point the time spent on the rest of the tests becomes wasteful fluff.

On that note, perhaps Tom's could determine the minimum fan speed required to hit and maintain the targeted max frequency? That seems a bit more useful to me, as it will tell me, the potential consumer, just what I'm going to have to endure, noise wise.

 

[DrKlahn]

Since quiet mode is an official mode from AMD, what's wrong with that? As I said before I agree with you that people might play in uber mode but this is AMD's fault for doing it from the begining, not Tom's Hardware who they found something wrong about it.

There's nothing wrong with it. It has been known since the first review that quiet mode is about sacrificing performance for acoustics. In this case vendor manufacturing choices caused the performance to vary more than expected. The point is that this is not the huge issue it was made out to be. Testing the card in its performance mode negates this issue. I don't expect a $550 graphics card to purposefully tested in it's lowest performing mode. Do you? The new C7 Corvette has an Eco mode to maximize fuel economy and a track mode to maximize handling and performance. Do you race the car at your local track in Eco mode because it's a valid mode? Of course not. It would be absurd. It's also absurd to do a performance evaluation on a flagship graphics card in a mode designed to specifically limit performance. AMD and it's partners are guilty of an oversight in quality control. Tom's motives aren't entirely innocent either. You only need to look at how the information was presented to deduce that.

So now the option is there but we shouldn't make it a big deal, if it isn't working properly...

Be honest here. Do you believe that this card is being bought by the majority of its owners for its performance or its acoustics. Be sure to take into account this has not been a issue reported on other sites or user forums. If you can honestly answer acoustics, then to YOU it would be a big deal. If you answered performance, which is essentially anyone I know that games on a PC, then it is not a big deal because you would not encounter it with higher fan speeds.

Who said "purposefully"? Accidents do happen though...
Would that change the fact that press cards and retail cards dont work the same at quiet mode?

Do you want me to the link the "Golden Sample" meme that Tom's saw fit to put into their article. That's a borderline accusation of fraud. They are insinuating that AMD purposefully sent the press cards that performed better than the retail samples are able to. The absolute performance of the retail cards was not compromised, but was being portrayed as such. I ask you again, do you feel that Tom's should have been testing a flagship card in its slowest mode against its competitor to begin with? Do you feel it was ethical to level these accusations initially with no hard data to explain them? Both of those things happened. I can't imagine any PC gamer assuming that a performance evaluation was being done with anything other than performance oriented settings.

 

[Leamon]

Another example of AMD basically lying to us. I swear they knew about crossfire issues, but did nothing to fix them until someone noticed. And now they're lying about the performance of their new card, someone noticed, and now they'll "fix" it. But they knew.

 

[DrKlahn]

The problem is solved, now they are the same at quite mode, read the article again, the variance of 3% had already been announced by AMD, and that wont chnage anything

The issue is indeed resolved. And we're all better for that. But I'm not giving Tom's a pass on how poorly this was presented. Or how absurd it is to be evaluating a $550 cards performance in it's non-performance mode to begin with.

 

[Gurg]

While I appreciate the hoopla Tom's is making about the issue, if everyone is really so concerned with getting to the bottom of this, why can't I find any charts showing the chip running under 100% fan load? It's not about noise at this point, it's about seeing if the chip can hit and maintain it's targeted max frequency with consistency, and if it can, it's pretty clear the cooling is insufficient, at which point the time spent on the rest of the tests becomes wasteful fluff.

On that note, perhaps Tom's could determine the minimum fan speed required to hit and maintain the targeted max frequency? That seems a bit more useful to me, as it will tell me, the potential consumer, just what I'm going to have to endure, noise wise.

I agree and also would like to see the testing done at both 78ºF as Tom's evidently does as well as ~70° room temps like my house in Denver. An eight degree difference in room temps could make a big difference especially when testing on an open system.

This must be one noisy little 92mm fan on this card as I actually sit next to my HAF912 case which has 5 120mms, a 200mm, an h100 and a 3 fanned windforce video card overclocked to the max. . .

 

[DrKlahn]

This must be one noisy little 92mm fan on this card as I actually sit next to my HAF912 case which has 5 120mms, a 200mm, an h100 and a 3 fanned windforce video card overclocked to the max. . .

Someone gets it. I always see the hoopla over acoustics yet all of my friends that PC game have at least half a dozen fans going in their cases. My NZXT Switch 810 has 9 120mm fans all whirring away. My reference 7970 is running in the mid 50 percent range on fan and I can just barely hear it with everything else moving air through the case and radiator.

I've been following the Overclockers.co.uk thread and it seems like 55% will keep it happy at stock in most all scenarios. If you want overclock or keep it from dipping even for an instant you start getting into the 60% range. I think some are setting the fan limit at 100% and seeing it max in the low 70% range for substantial overclocks. Noise is reported to be comparable to the 79xx fans.

 

[littleleo]

what do you mean Novuak?

If you have an issue here's a tissue.

 

[Memnarchon]

I am tired of this, DrKlahn.

Quiet or Silent mode is an AMD feature that doesn't work out of the box as the press card did. Tom's Hardware discovered it and made an article about it.

If you have any evidence that Tom's Hardware did the article in order to hurt on purpose AMD sales illegally, you should take the evidence to the DA's office.

 

[yyk71200]

Another example of AMD basically lying to us. I swear they knew about crossfire issues, but did nothing to fix them until someone noticed. And now they're lying about the performance of their new card, someone noticed, and now they'll "fix" it. But they knew.

The issue is with 3rd party cooler, not with AMD's card itself.

 

[md1032]

Really great article. I can always count on Tom's Hardware to show me some real data while the boys are talking theory on the forums.

 

[GOM3RPLY3R]

I'm going to say that the fact that the card runs so hot to get performance that is ~15-20% better than the Titan, and really on any cooler (for some reason) having a huge noise from the card once you play a game, shows that AMD really hasn't done anything except throw more crap into a card and clock it higher.

The full capabilities of this card will definitely be achieved using water cooling solutions, but even then, you'll be looking at near, if not more than Titan prices to get the cooling you need.

Otherwise, I'm not sure why people are concerned about heat. I had an old 5570 where without me realizing it, the GPU fan worn down to the point that it was only running ~200-300 RPM. The card ran for about 2 hours at 100-115ºC. Ran my games fairly well and it still works perfectly fine to this day. Unless AMD went all "dainty" on their GPUs, you can just set the Curve lower to reduce the "noise" (it doesn't bother me), and Let it run at 100-105ºC.

As another solution, the ASUS GTX 770 DCuII is going for $340 right now off of newegg. Get two of those bad boys for ~ the same price as it, except never go above 70ºC and get ~30-40% better performance. (Just a recommendation ^_^)

Just to let you know, I'm getting a max of about 64ºC with the Core at 1260 MHz and the Memory at 7500 MHz and the stock fan speed. Another note, even at its 3800 RPM max, with all my case fans turned off, I can only hear a small push of air when I have me headsets off and ear near the case. I'm telling you guys, even if you are a AMD guy/gal, this is something you may not want to miss out on. The 780 is even only $520 and has ~ Titan performance. The ASUS DCuII version has better than Titan performance when you set the clock at ~1050, which is fairly easy and simple.

UPDATE: Apologies, the GTX 780 DCuII is at $499. ^_^

 

[danwat1234]

Why is the fan locked to 40% PWM? Why can't it go to 100% for uber cooling and full clock speeds?

 

[ferooxidan]

It is funny if all the owner of r9-290 x or not didn't point out or found this problem. Here is one simple question: DID THEY OWN A PRESS SAMPLE CARD SENT BY AMD FOR THEIR COMPARISON?

How did they make claim anyway if they only have one and had no such comparison?

 

[megamanx00]

Shame on you AMD. Shame. Too much promise, too big of a GPU die, and an inadequate cooler. It just seems that all the changes since letting go of Dirk Meyer have lead to a less competitive and less honest company.