AMD Radeon R9 290 4 GB

[PopcornMachine]

There is an article on the store samples at Tech Report: http://techreport.com/news/25609/up...9-290x-cards-may-be-slower-than-press-samples

One poster there showed his card throttled down to ~750MHz until he upped the fan speed to 50%.

Probably due to local case/envrionment, or it may be that some chips are more sensitive to the 95C temp.

If I get a 290, and I prob will, it will get a water block.

 

[Eroticus]

There is an article on the store samples at Tech Report: http://techreport.com/news/25609/up...9-290x-cards-may-be-slower-than-press-samples

One poster there showed his card throttled down to ~750MHz until he upped the fan speed to 50%.

Probably due to local case/envrionment, or it may be that some chips are more sensitive to the 95C temp.

If I get a 290, and I prob will, it will get a water block.

"More likely is that whoever got a bad results didn't have enough cooling and that the card went into thermal throttling."

Variations exist, but does not affect everyone
With another day of data at hand, it is clear that differences between graphics cards available. Specimens from PowerColor include a slightly slower fan combined with a slightly higher voltage to the GPU, which together provide lower frequencies as the temperature puts a stop.

This, however, only limited Quiet mode where the fan speed is locked to a maximum of 40 percent. In more Über released, which is what SweClockers use reviews, the difference in performance nonexistent - both press copy and PowerColor card can operate at peak frequency even for long periods.

AMD has a solution in the offing
At AMD, work is very hard on the problem, where jobs weary European Chamber of writing excites newly awakened colleagues in Canada. The company has a solution on the table, and promises a driver update within 24 hours to minimize the problem.

We will be releasing a driver in the next 24 hours That corrects this behavior by Normalizing all fan behavior. The 290X in Quiet mode Should be at 2200rpm, and the 290 Should be at 2650rpm.

In short, the company will be using software to normalize the different cards fan speed. In the current situation this is controlled only by a percentage, but after the update will also actual RPM value is taken into account. The company hopes that this will even out the differences between different copies, and fit while on the ground for higher performance.

For Radeon R9 290X the value 2200 RPM for Quiet Mode, compared with about 1,920 RPM SweClockers PowerColor-copies and 2,020 RPM for press copy. That means a potential prestandabump for both models - of course at the expense of slightly higher noise level. Performance Mode Über remain unchanged.

For scaled-down Radeon R9 290 added value to 2650 RPM, compared with about 2,500 RPM SweClockers press copies. Even where it is a question of a higher value, which in theory means more performance. It is worth mentioning, however, that the fan speed for Radeon R9 290 already suffice quite well.

All AMD coolers always was suck ... the only normal cooler that AMD made it was on 7990 that all.

 

[manofthem]

if I get a 290, and I prob will, it will get a water block.

Do it, do it, do it

 

[mastershake575]

Few things I noticed

1. Third party cooler versions should be out in the new couple of weeks so I don't know why people are jumping ship. I would rather wait 3-5 weeks and pay only $25-50 more to get better temps and a guaranteed 1050-1100mhz overclock (A third party 290 at $425-450 clocked at 1050-1100mhz while having good temps seems like a no brainer and I would for sure wait for that).

2. The price tags ($400) gets me excited since I don't have to upgrade till 20nm comes out (middle to late of next year). Either the 290 or the 20nm equivalent should easily be $300 come late next year (if crossfire users are having good results, I might just have to buy two at that price)

 

[The Von Matrices]

A Swedish magazine already debunked this. They bought over the counter samples and tested against review samples and they were just a fast. Absolutely no difference.

More likely is that whoever got a bad results didn't have enough cooling and that the card went into thermal throttling.

http://www.sweclockers.com/artikel/...der-for-golden-samples-sweclockers-undersoker

So you haven't read the article you link to? You are not aware of the extensive update on that page?

I don't know if anyone else read the entire article, but specifically talks about AMD making the 290X louder in the next driver update to be released tomorrow. To quote (translated):

At AMD, work is reportedly on the problem, where jobs weary European Chamber of writing excites newly awakened colleagues in Canada. The company has a solution on the table, and promises a driver update within 24 hours as by his own admission to minimize the variances.

We will be releasing a driver in the next 24 hours That corrects this behavior by Normalizing all fan behavior. The 290X in Quiet mode Should be at 2200rpm, and the 290 Should be at 2650rpm.

In short, the company will be using software to normalize the different cards fan speed. In the current situation this is controlled only by a percentage, but after the update will also actual RPM value is taken into account. The company hopes that this will even out the differences between different copies, and fit while on the ground for higher performance.

For Radeon R9 290X the value 2200 RPM during Quiet Mode, compared with about 1,920 RPM SweClockers PowerColor-copies and 2,020 RPM for press copy. That means a potential prestandabump for both models - of course at the expense of slightly higher noise level. Performance Mode Über remain unchanged.

For scaled-down Radeon R9 290 added value to 2650 RPM , compared with about 2,500 RPM SweClockers press copies. Even where it is a question of a higher value, which in theory means more performance. It is worth mentioning, however, that the fan speed for Radeon R9 290 already enough relatively well.

Their press copy had too low a fan speed, and that resulted in less performance in quiet mode but less noise (the fan speed on their press card was 10% lower than the setpoint, and the retail card was 13% lower than the setpoint) That means that for many users the driver update will make their card louder than they already were before the driver update. 10-13% may not seem like much but it is certainly not welcome on an already loud card.

Wizzard, when the update ships, could you retest the 290X performance and noise in Quiet mode and the noise level in Uber mode? They may be higher than when you tested them originally.

 

[Am*]

All I can say is thanks AMD. Now we have $500 performance in a $400 card, minus 5% here or there.


Your move Nvidia; this card just demolished the value of your entire mid-high end lineup -- hope you're prepared for another wave of price cuts, because in another month or so you won't have a hope in hell of charging a $20 premium against any AMD card of similar performance, let alone $100. 780 Ti hits tomorrow and we'll see if it is of any value to anyone, because from what I've read so far ($699 price for a 3GB SKU), it looks like another worthless e-peen card not worth jackshit to anyone. If it is as overpriced as I think it will be, I will more than likely be switching camps this time -- even if I have to put my 3D Vision crap aside for another while.

 

[Fluffmeister]

nVidia cards have been selling just fine with a hefty price premium for donkeys already, even with AMD bundling countless games too.

Now AMD... stick a decent cooler on your cards, thanks.

 

[radrok]

nVidia cards have been selling just fine with a hefty price premium for donkeys already, even with AMD bundling countless games too.

That's because there was a huge performance gap before the 290x/290.

 

[dom99]

Got my 290 today, can't wait to stretch it's legs, I'm aiming for 1.1ghz on stock voltages like I've seen quite a few others hitting

 

[Casecutter]

I'm not so sure about your timeline being reasonable. GTX 680 & GK104 were released in March 2012 and shocked AMD with its performance (it shocked NVidia too). Yes, AMD released the 7970 GHz edition to compete, but surely they had to know that NVidia still had GK110 to release and that Tahiti wasn't going to be able to compete with it.

I don't think AMD really thought a GK110 on the gaming market as something they saw as reality until say mid Q3 2012. When Tesla cards started to come to light and showed the promise of great efficacy. I think AMD knew it might take time to amass binned chips and what exactly their yield would look like, would they use chip with 12-14 SMX’s in less than just the professional Sku's? If they had to use part gelded even more would the perf/watt look good? AMD was look forward to end of 2013 with a 20Nm...

Then as Mathragh said, both shoes' dropped near the end of 2012! TSMC communicates that 20Nm risk production is pushed back into end of Q1 2014 (now some figure later); Nvidia got the same memo but Kepler wasn't as effected. Titan (14-SMX) was be pushed forward to show in 3 months, while the amassed a war-chest of 12-SMX parts that didn't provide the efficiency imposed for the Tesla professional products.

Though AMD needed something they could rebrand, but need a true Enthusiast part to work the cascading product stack they had planned. So the whole "we won’t be releasing any new cards for the bulk of 2013" went out. The other thing was Titan seemed overwhelming at first, but 25% more performance for $600 wasn’t that intimidating. AMD was already in the design scaling the 20Nm layout to spin on 28Nm production.

http://www.techpowerup.com/180009/no-new-gpus-from-amd-for-the-bulk-of-2013.html

 

[Recus]

So you haven't read the article you link to? You are not aware of the extensive update on that page?

He thinks that problem denying fix the problem.

http://i290.photobucket.com/albums/ll264/ahrinin/MotherOfGod.jpg

All I can say is thanks AMD. Now we have $500 performance in a $400 card, minus 5% here or there.


Your move Nvidia; this card just demolished the value of your entire mid-high end lineup -- hope you're prepared for another wave of price cuts, because in another month or so you won't have a hope in hell of charging a $20 premium against any AMD card of similar performance, let alone $100. 780 Ti hits tomorrow and we'll see if it is of any value to anyone, because from what I've read so far ($699 price for a 3GB SKU), it looks like another worthless e-peen card not worth jackshit to anyone. If it is as overpriced as I think it will be, I will more than likely be switching camps this time -- even if I have to put my 3D Vision crap aside for another while.

PMS again.

 

[Tatty_Two]

9-15% actually but effective well.

So when non reference coolers do become available that becomes very realistic compared with say the performance difference between the Titan and 780.

 

[mhadina]

botlleneck

How stupid would be to pair it wit the overclocked AMD 4-core APU and 8 gigs of ram ?

 

[Mussels]

How stupid would be to pair it wit the overclocked AMD 4-core APU and 8 gigs of ram ?

i dont think you'd get 100% out of the GPU, but it would still be plenty fast.

 

[Tatty_Two]

How stupid would be to pair it wit the overclocked AMD 4-core APU and 8 gigs of ram ?

Dependant on which APU, some will overclock to 4.5Gig+ so there should not be any bottleneck there, but at stock possibly, but to be fair, some lower 4 core Intel offerings that stock at much lower speeds than the A8's and A10's, despite being faster clock for clock would also be a bottleneck to some degree.

 

[jormungand]

Your post reminds me of someone who buys a sports car but can't afford to drive it.

Ive been driving $400+ cards long time ago, i always wait for the coolest and best performance card and at the end of the war i make my choice, what i stated is that the non reference cards maybe come better made for all interests. IMO!!!!

 

[Am*]

PMS again.

I know, I know -- it's your time of the month yet again, so by all means continue to pointlessly quote other people's posts and forget the fact that this is a USER FORUM with people posting their opinions on stuff -- which you can ignore, debate against or kindly GTFO. You're excused for this week, or until your tampon tantrum blows over, whichever comes first.

 

[ChristTheGreat]

I don't know if it has been posted, but: http://www.techspot.com/review/736-amd-radeon-r9-290/page8.html

I took the IceQ X2 cooler off the HIS Radeon R9 280X and stuck it on our R9 290 sample. Cooling was dramatically improved. The FurMark stress test maxed out at 76 degrees while the card never exceeded 63 degrees in Crysis 3 and Battlefield 4. So it seems as expected the board partners will be able to solve the heat issues of the reference card.

it's interresting.

 

[Mathragh]

Another way to go might be :

apparently someone at techreport did this, and had no more throttling problems, even with lower noise. lol.

Though one has to wonder if its worth the lost warranty, and whether there isn't a way of achieving this more cleanly, like getting rid of the whole bracket.

 

[RCoon]

How stupid would be to pair it wit the overclocked AMD 4-core APU and 8 gigs of ram ?

<- see system specs + 8GB extra RAM.
1440p and 0 complaints.

Another way to go might be :
http://i.imgur.com/aSzp66V.jpg



apparently someone at techreport did this, and had no more throttling problems, even with lower noise. lol.

Though one has to wonder if its worth the lost warranty, and whether there isn't a way of achieving this more cleanly, like getting rid of the whole bracket.

Link? I'd love to know just how much thermal improvement there was because of this.

 

[Mathragh]

Link? I'd love to know just how much thermal improvement there was because of this.

He didn't post benchies, just made the claim that it solved all his crossfire problems. Would also like to see some benchies tbh lol.
However I suspect that its a bit too late for any comparison between the former state and current state .

 

[HTC]

Another way to go might be :
http://i.imgur.com/aSzp66V.jpg



apparently someone at techreport did this, and had no more throttling problems, even with lower noise. lol.

Though one has to wonder if its worth the lost warranty, and whether there isn't a way of achieving this more cleanly, like getting rid of the whole bracket.

Quick: someone SEND THIS INFO to AMD


I did similar to my Obsidian 650D case by cutting open the front vent mesh that came with it as, even with the fix that Corsair sent me, the noise was very noticeable for my liking: this way, the noise went down quite a bit to much more tolerable levels.

 

[Steevo]

Another way to go might be :
http://i.imgur.com/aSzp66V.jpg



apparently someone at techreport did this, and had no more throttling problems, even with lower noise. lol.

Though one has to wonder if its worth the lost warranty, and whether there isn't a way of achieving this more cleanly, like getting rid of the whole bracket.

I have never liked the AMD/ATI EMI shield plate. Perhaps someone should do a completely open one with carbon fiber to improve the look and flow

 

[erocker]

Another way to go might be :
http://i.imgur.com/aSzp66V.jpg



apparently someone at techreport did this, and had no more throttling problems, even with lower noise. lol.

Though one has to wonder if its worth the lost warranty, and whether there isn't a way of achieving this more cleanly, like getting rid of the whole bracket.

That's not a bad idea. I think it could of been done much cleaner if they took a little more time. It will be interesting to see in anyone else does this.

 

[N3M3515]

About the noise catastrophe everyone is talking about, my own vcard on the review here in tpu is rate 48db at full load, that's 1 point less than R9 290 and 2 less than the X variant.
I honestly can't even hear the vcard's fan when playing...
link: http://www.techpowerup.com/reviews/MSI/HD_7870_HAWK/27.html