AMD Radeon R9 290 Review: Fast And $400, But Is It Consistent?

Page 6 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Status
Not open for further replies.

somebodyspecial

Honorable
Sep 20, 2012
1,459
0
11,310


But since AMD admitted they have "variance" problems and are working on a fix, Toms is basically vindicated right? They may not have had time to do anything more (they may not even own the retails themselves). I posted a link to TR forums showing another guy having this issue, and he replaced the grease, ran with the side off etc. Nothing changed until setting the fan at 50%.

I see nothing wrong with reporting what they found. It's just the data they got while testing. How do we know an article isn't coming already? Techreport is planning on testing retails for issues (due to the toms article) so look for that soon if you don't get it here. Though again, AMD has admitted to a problem, so there isn't a question about that now. It's a KNOWN fact.
 


That doesn't matter, remember the GSOD issue? The party faithful still refuse to acknowledge that for crying out loud! :lol:
 

DrKlahn

Honorable
Nov 5, 2013
24
0
10,510


As I pointed out above, the TechReport issue seems to be a misunderstanding or unrealistic expectation on the part of the card owner. It is well documented in all the 290x reviews that the Quiet profile with its 40% fan speed cap will throttle the clockspeed. The Uber mode with its 55% cap is intended to allow the card to run without throttling. So I'm not sure why that person thought the card would run consistently at 1GHz with less than 50% fan speed.

I've never stated they should not report on this potential issue. However they provide very little data and infer some serious accusations based on it. There should have been a separate article which looks at the throttling vs. clockspeed. And any potential for assembly/QC issues. A quick browse through the enthusiasts forums would show that this is not the norm for these cards, which should have prompted more caution about presenting the information.

AMD has stated it believes there is a problem with those particular cards. This does not indicate or admit that there is any widespread issue or design flaw. Tom's, a site I have read for a very long time, is not vindicated. It's a rushed "conclusion" with very little data. Several sites are already picking this up as a real issue. It's essentially sensationalist journalism at this point until we see some real investigating. Personally I expect more out of Tom's.
 


And if they had said nothing then you would just slag them off later on, it's a no win situation and not their fault that you don't like facts that don't blow the smoke up AMD's arse.
 

Alexfitz

Honorable
Nov 6, 2013
2
0
10,510
I will wait on the Windforce edition that of the 290. I love that cooling system and it is very close to the kit the author details here. Hopefully that version of the card will be out shortly
 




Can not argue with that Mouse. Hell if this is how the cards released then that is what they report on.

Tom's CLEARLY stated that this review is preliminary,
 

DrKlahn

Honorable
Nov 5, 2013
24
0
10,510


Not correct at all. I would have put in a caveat that they are seeing some performance problems with retails boards and that a full report is forthcoming. Instead we have people jumping to conclusions and getting out their pitchforks for an issue we know very little about. Other sites are picking this up as a real problem.

What facts do we know? Tom's acquired a retail card. The card shows throttling and goes above 40% fan speed. We also have several pages on other enthusiasts sites with reporting normal behavior in regards to clockspeed/fan. These people are posting clockspeeds, screenshots, benchmark ratings, and in game performance. So as a journalist do you infer that a major corporation is behaving in a fraudulent manner based on your findings from a very limited sample? Or do you take a step back and say you are seeing anomalies and that you will make a full report after thorough testing and further investigation? I know which way I would go.
 


Knowing of the large amount of AMD PR shills that are out there skewing things I think the right call was made.
 

lowstandards

Distinguished
Feb 10, 2004
90
0
18,630
I 100% agree with this article on the page with the Arctic Accelero Extreme III aftermarket cooler. Let me explain:

I have a 290x and with the stock cooler the noise that beast would put out was unacceptable! The GPU, although designed to run at 94 degrees, reaches that 94 degree mark very quickly in games and benchmarking. Very quickly! Almost like it was designed to do so.

Anyway, 2 days of this howling in my Corsair 800D case was long enough so I purchased the Arctic Accelero Extreme III aftermarket cooler and....ZOINKS!!! This is what should have shipped on the 290 series. In Catalyst Control Center, I set the Fan speed to "up to 100%" and target GPU temp to 70 degrees. No overclock. With these settings the fan is literally barely audible and...get this, GPU temp maxes out at 69 degrees and the fan hits 55%! What?!

So my 290x is running 25 degrees cooler and is, what I consider silent, and I know there is room from some overclocking. I prolly won't overclock though since I only game at 1920x1080 all settings maxed and my games run silky smooth.

Seriously, either wait for AMD partners to release 290 cards with aftermarket coolers or buy a 290 now and get the Arctic Accelero Extreme III. You will NOT regret it.
 

lowstandards

Distinguished
Feb 10, 2004
90
0
18,630
The Arctic's Accelero Xtreme III needs no modification for the 290 series.It comes with all the heatsinks needed as well. Heck I had a LOT of left over heatsinks. I have the 290x but I read that the Accelero cooler is compatible with the 290 as well.



No modification is necessary. I did run out of sinks for the VRAM but just paired up some leftovers and used those.
I don't regulate the voltage, I rely on PWM. In the drivers I set the MAX GPU temp to 70 degrees; MAX FAN speed to 100%. By forcing the GPU temp to max out at 70, this in turn forces the fans to speed up earlier and quicker. I chose this so that I can get a higher fan speed sooner so that the chips on the board get ample cooling sooner. Recall that the air has to flow through the main heatsink to reach the chips on the board. So far this has worked great for me! And even with the fans hitting 80% fan speed (in Crysis 3 after 3 hours of gaming), it is still LITERALLY 50+% quieter than the AMD stock cooling solution. Embarassing for AMD.

P.S. I am overclocked on my GPU by 10% and 15% on memory.
 

Crashman

Polypheme
Former Staff
I think you need a dictionary, you appear to be confusing the word "instigated" with "investigated, and the word "myth" with "fact". Chris spent the money to investigate and accurately reported his findings. Which, in turn, makes your analysis of his findings the only portion of your post that was truly biased. I'm not sure about you, but most of us like the truth :)

What the heck are you talking about you...you...you...MOUSEMONKEY!

 

uborapnik

Honorable
Nov 6, 2013
2
0
10,510
Can't wait for the guide to installing Xtreme III cooler. Also, if you can, tell the difference between 290 and 290x since some of us will be installing it to 290x. :)
 


Scrumworks knows... ;)
 

natoco

Distinguished
May 3, 2011
82
0
18,630
Reminds me of the gtx 480, still a great card, but a hot and power hungry one at that. They have overclocked there silicon right to the very limit and then let the software just turn it down. At least with an aftermarket cooler it would be better but I doubt much tweaking from them. Let the cooler wars begin and see what the board manufactures can pull out of the hat. Ps with electricity costs quadrupling in the last 3 years in Australia and extra 30 watts per hr and say 6hrs a day is NOT cheap anymore, these things need energy star ratings with yearly running cost stickers on them.
 

slomo4sho

Distinguished


Just for you:

instigate
verb (used with object), instigated, instigating.
1.to cause by incitement; foment
2. to urge, provoke, or incite to some action or course

myth
noun
1.a traditional or legendary story, usually concerning some being or hero or event, with or without a determinable basis of fact or a natural explanation, especially one that is concerned with deities or demigods and explains some practice, rite, or phenomenon of nature.
2.stories or matter of this kind
3.any invented story, idea, or concept:
4.an imaginary or fictitious thing or person.
5.an unproved or false collective belief that is used to justify a social institution.

The words fit perfectly fine within the context of the post... maybe you just need to expand your vocabulary ;)

Jokes aside, the so called "test" was of two aftermarket cards for which no information was provided and were likely from the same merchant and manufacturer. Chris made a clear accusation on the last page of the review and even made sure to bold it so no one could possibly miss the alarming statement that "Clearly there’s something wrong" with the 290X retail products and that the 290 will likely follow suit. How can you call this objective journalism? Based on 2 card? Based on no real details? Since you can't seem to understand how such claims can rouse the ignorant masses to run with such a claim then it is obvious why you seem to think that Chris's rant was factually based and involved investigative reporting. Oh look, now GURU3D also invalidated Chris's claim which actually seems to have originated from Nvidia. So that is what, two review sites now that have disproved your "fact"s and "truth"? I am sure there will be more to follow. Did Chris forget about his own review of the 290X?

If the card is running in Quiet mode, hits 95 degrees, and cannot control temperature under 40% fan speed, it’ll start pulling back clock rate to avoid 96 degrees. Performance takes a hit in the interest of low noise. Switching to Uber mode simply gives you 15% more duty cycle before clock rates start dialing back.
clock-rate.png


I debated about where to put this graph. In one sense, it belongs with my CrossFire data because it shows that heat hurts the way two R9 290Xes perform. But I'm putting it here because this is an illustration of PowerTune in action. The technology, for better or worse, is forcing these cards to abide a 40% fan speed. So, when the GPU hits 95 degrees and can't spin its fan any faster, you have to watch the core clock melt away. The effect is even more severe with two cards next to each other (even with space between). Hawaii is still a very fast GPU, in spite of this phenomenon, but it's a shame to observe, regardless.

You’re certainly free to manually specify higher maximum fan speeds than the 40% I used, but it’s pretty telling that even AMD’s Uber mode stops at 55%. Again, we’re dealing with a reference cooler that makes a lot of noise once it gets going. I’d personally leave the card set to its Quiet firmware in my own PC.


I also like how you took a portion of my post and declared it biased by ignoring the full context of the text. Great job :)

Also, the GTX 780 Ti was launched today, did you guys not receive a sample to review?

MSI_GTX780Ti_specs.jpg




There is much irony in your name :)
 

somebodyspecial

Honorable
Sep 20, 2012
1,459
0
11,310
Well toms has this in the article (now? did I miss is the first time?) for everyone whining:
"Update: As is Tom's Hardware policy, we shared these potentially problematic findings with AMD prior to publication, and the company insists something is wrong with the retail-purchased cards I tested. We will continue investigating and, if any additional news becomes available, update this story"
http://www.tomshardware.com/reviews/radeon-r9-290-review-benchmark,3659-2.html

Also it looks like NV set the 780 TI just below the noise/watts of 290x...LOL. Shocked ;) So they've figured out two can play the same game it seems :) But hey, is that 960 DP cores real? Titan has 896, so this means they just made a BETTER TITAN also and selling at ~700? So they are taking titan off market or something and just having one to rule them all?
http://videocardz.com/47719/nvidia-launches-geforce-gtx-780-ti

Running through that NON-review now, it looks like they only matched quiet 290x but not really clear how fast it was running for 290x, while uber would be worse. 1274 for OC though, "Just as I told you, GTX 780 Ti is loves overclocking. Here’s an example with 1274 MHz clock, more than 250 MHz additional power." They didn't need water to get there either ;)

IF the DP and everything is correct, with 2880 cores (full SMX) this card just killed Titan in ALL areas and will no doubt fly off the shelves. We only have 4 more hours before all reviews start hitting. I'd be lying if I said I'm not excited here...LOL. But not to buy this, just what the heck will they have to do for a top 20nm to beat this now to be worth buying it? MONSTERS are coming from both sides next year due to the apparent pissing contest going on right now between AMD/NV...ROFL. I'd price this thing $800. King in everything and replacing the $1000 DP power by a lot if they all work for PRO people. Even with default clocks this will beat a Tesla (which runs slower with same 2880 etc). Even at $800 that's 1700 off of Tesla and you get top gaming to boot. No point in Titan if this has better than Titan's DP FP.

Now they just need to release a new driver for 780/titan owners that upclocks to the same noise/watt level. Or two drivers, one for the noise haters (current driver) and one that ups to more noise/heat/clocks like it appears they've done with 780TI (sorry Chris A. they followed AMD down the rabbit hole). I'd say that issue bugs me, but they have set the perf bar higher for 20nm now by doing it so I don't care (I'm not buying one, waiting for 20nm). Seems they priced it $700 so very good deal if shipping with 3 AAA games and $100 off shield still. Note SLI barely raised the DB, which lands at 1db over AMD 290x. On purpose or accident? Will we see a vid now on NV's youtube page showing SLI 780TI's running at the same noise as ONE 290x? Dirty but as a business I'd do it...LOL.
 

yannigr

Distinguished
Oct 3, 2008
140
0
18,680
Total BS about press samples been golden samples. That's what I am reading today and you are one step from deletion from my bookmarks after almost 15 years that I read you every day.
 


Wow, that rage. That fanboyism. Them threats!
 

somebodyspecial

Honorable
Sep 20, 2012
1,459
0
11,310


There is much irony in your name :)
[/quotemsg]

I don't think so... :)
But if it makes you feel any better, in a previous toms post (response to some other joker :)) I made the comment "NobodySpecial" which I use almost everywhere else was taken here...LOL. Read up to 500 posts in your thread...still nothing useful to show toms is wrong and now we have AMD saying there is a problem which will be fixed with drivers or something.

Done reading that OC thread for a while, will wait for you to post some post#'s showing similar cases with different results than toms had (as noted before, not with water etc..but running as toms did on retails). No time to read the other 2500 posts in your thread for now. It's up to you to prove there is actually anything in there useful to prove toms wrong. :) Especially after accusing bias, and AMD admitting to problems that will be supposedly fixed by a future driver. AMD is admitting there is at least SOME credibility in what toms has showed.
 

somebodyspecial

Honorable
Sep 20, 2012
1,459
0
11,310
I think it is time for you to calm down. Because we talk gaming here and Titan is in the chart for comparison's sake. Why? Because Titan launched as a flagship gaming card that can compute. GTX 780 is the little brother of Titan, until the Ti arrives. Kudos to Nvidia for making a card that is good in both compute and gaming. It is something that deserves appreciation. However, the point here is not that. R9 290/X is INDEED a Titan slayer due to its price AND gaming performance combined, if we are looking from this perspective. In this view, prosumer features of Titan aren't so relevant to the gamer. But, for the fact that Titan was launched as a gaming flagship, it is in the chart for comparison's sake. There are probably people that bought Titan just for gaming purpose, as you said earlier.

What in my comments says I'm not calm? All the LOL's, Grins, ROFL's mean I'm angry or not calm? :) I thought all that said the exact opposite. Upset I know how to prove my point with data and reasonable arguments backing it? This is a discussion forum...Discussing an article is exactly what we're here to do right? The article has DP etc benchmarks so not out of line discussing the NON gaming side of these cards whether you like it or not. Tell toms (and every website, since they all do it) that nobody cares about compute/GPGPU etc and see how that flies ;) Titan is in the charts because it is the former king, not just for comparison. In a review of top cards you can't leave out what was top just a few weeks ago. Yes as I said there are "probably" people who do buy for JUST gaming. But I guarantee you there are a TON of people buying it for MORE than gaming (and see $1000 really cheap vs. Tesla/Cuda) which is exactly what I explained.

Never said it wasn't a TITAN SLAYER for people who are GAMERS ONLY with no pro ambitions at all. But for the rest and the majority of who's buying them? Titan is the steal for a lot of DP work/pro stuff, though I think the 780ti just made it pointless if it really has fully working 960 DP cores and faster speeds. That would make 780TI a Titan slayer for EVERY user, not just gamers...and a 290x slayer too... :) 25% more cuda cores than 780, faster clocks, faster mem. Very nice. Should we call it the everything slayer? It's already been OC'd to 1275.

Edit, messed up the quote... ;)
 

drezzz

Honorable
Oct 6, 2013
13
0
10,510
AMD release 290X and NVidia drop price on 780. Now AMD release 290 and it still beats the 780, almost matches the 290X, and costs $400. Nvidia are dead.
 

mgae

Honorable
Sep 27, 2012
7
0
10,510
On the review from techspot here http://www.techspot.com/review/736-amd-radeon-r9-290/
, they replace reference cooler with His IceQ X2 from R9 280X and temperature drop to 60-70 immediately. Imagine R9 290X with that cooler, there should be no throttling and about 10% performance gain.
 

fanman

Honorable
Oct 9, 2013
10
0
10,510
We have 2 NV-salesmen here, and if AMD succeed in making them miss their target sales they will lose all the bonus. Its November now ;)
 

technoholic

Distinguished
Feb 27, 2008
800
0
19,160


Unless it's built on a new lithography (which is unlikely), all these added performance equals to more power draw = heat, i'd think :D Yeah, we are waiting for the benchies today
 
Status
Not open for further replies.

TRENDING THREADS