This card needs a better cooler for anything but short benchmarking runs it seems to me.
It throttles heavily if left at the stock fan setting, and even with the "Uber" jet-engine setting it throttles still.
http://www.techpowerup.com/reviews/AMD/R9_290X/30.html
I'm sure it would be lovely on water or with an improved cooler
For those guys worried about throttling I just did some extra testing over the past day. I ran the Crysis 3, Max Payne 3, Metro Last Light, Bioshock and Tomb Raider tests over and over again for an hour and then recorded the results. The frames per second in all those games went unchanged from our original results.
The core frequency was monitored and never fluctuated from 1000MHz though temperatures in all games were between 92 – 95 degrees and the card was quite loud. Testing was conducted in the Uber mode.
I am not sure what game TechPowerUp tested in but I cannot reproduce those results.
This gave me a pretty good laugh. 2000 million more huh?
Also on one of the 99th percentile graphs it said "12 seconds between frames" which I am sure meant to be 12 ms between frames.
I am only going to say this once more, a billion is not measured the same way globally. A billion can either be 1000 million or a million million, kind of a big difference.
Frame times was better in quite a few of the tests, infact the only time the frametime was bad with SLI was with metro (if I remember right). But as far as the frametime argument goes the data in the review does not back up your claim.
Really, you might want to take a proper look. Overall the R9 290X was 22% faster when measuring frame time performance. The R9 290X was 12% faster in Crysis 3, 53% faster in Medal of Honor, 17% faster than Resident Evil, 45% faster in Bioshock, 54% faster in Metro Last Light and 127% faster in Sleeping Dogs. Are we reading the same review?
SLI not working? I'm sorry but when was the last time you played something where SLI was not supported or working? I've had two 560 Ti's for more than a year and never once had a single hiccup on any game ever with that setup. And to prove my point even further I got a 670 as a present and its performance is identical to what my two 560's were.
You must only play AAA titles then. Still the frame time results prove that even when SLI is working it’s not always working that well. In many of the games above the R9 290X will provide a much smoother gaming experience, even when the frames per second are slightly slower.
Also my point with the $500 thing wasn't very clear, what I meant was: the article goes on about how the titan is $1000 and this new R290 is supposed to be around $550 the 780 is damn expensive also. So if you are going to spend $500-550 then the best value for money according to the tests done was to get a pair of 660's. I don't see how you could possibly justify getting a 780 or even this 290X when SLI has proven to be very reliable and cost effective.
Your point still must not be very clear because you are wrong. If you are spending $500 - $550 the best option is the Radeon R9 290X, not the GTX 660 Ti SLI cards. If you buy the SLI setup today rather than the R9 290X you would have made a mistake.
Lastly the part about midrange dual gpu cards not making sense, do you seriously believe that?
1. Cooling, its far easier to handle two mild heat sources than it is to handle one insanely hot (and very small) one.
2. Performance, refer to the benchmarks in the article.
3. Price, cheaper than or the same cost as a single slot card of similar performance.
4. It has been done already by evga in the past with a dual gpu 560 I think but by the time the released that card it was many months after the initial release and it was substantially more expensive than two individual cards. I firmly believe if they released it faster and at a better price it would have been quite popular.
Yes I seriously believe that.
1. No, plain wrong. How can you possibly come to that conclusion when a pair of GTX 660 Ti SLI cards consume roughly 30% more power?
2. Yes the frames per second performance is a fraction better, the R9 290X was 2% slower overall. If you believe that is enough to warrant a dual-GPU setup over a single GPU then good luck to you.
3. Yes you are making my point with your third point, I agree.
4. Yes that is why the eVGA card failed. Its funny how you just don’t see dual-GPU GTX 660 Ti cards, what gives? It's such a brilliant idea.