LinkBack URL
About LinkBacksRead more.However it admits it is working on software / hardware performance variability issues.
Read more.However it admits it is working on software / hardware performance variability issues.
Interestingly the guy in the tech report forums had his card drop to around 850ish average speed in normal mode, but only needed to bump the fan speed to 50% (i.e. significantly less than the 55% mandated by "uber" mode) to get a rock solid 1GHz.
Anyone want to take bets that AMDs driver solution will be to force the 290X's fan speed to 47%, just like the 290's?
For crying out loud AMD, just let your partners loose on it and have done!
http://forums.hexus.net/graphics-car...ml#post3098097
Originally Posted by CAT-THE-FIFTH
Gibbo walked into the warehouse and pulled 2 random cards - both were better than the review sample!
Seems NV really are getting desperate now.
hexus - why not go to scan and get a random card and test it?
Probably needs a reference on that (i.e. that "Gibbo" is overclockers.co.uk's MD)
http://forums.overclockers.co.uk/sho...73&postcount=9
Not really sure what NVidia have to do with this story, they weren't the ones to "accuse" AMD of this AFAIK.
This won't take long to fix with a more aggressive fan profile anyway.
Fantastic - one review site buys one retail card which turns out to be a poor sample, and essentially accuse the entire brand of fixing benchmarks somehow. And of course, the sensationalistic headline spreads like wildfire across the web, seemingly without anyone attempting to check its validity.
Kudos to Gibbo for doing some unbiased testing on multiple samples!!
Yeah, sensationalism at it's best. Come on Hexus, update the article with other sources giving a fairer picture.
Surely the story is that there is *sometimes* some variability, for better or worse, from the review samples and between cards... but that is kind of to be expected, surely GPUs just like CPUs have some variation in how well they tolerate voltage/clocks and some just don't do quite as well, some cards heatsinks will be making slightly better contact than others or got a well lubed fan that manages a few more rpms, perhaps the ambient room temperature changed and the that just nudged the card past its cooling ability (seems like the 290 series is right on the edge of this)... good science is always done with many repetitions of many samples ;-)
The important part of the whole 290(X) debacle is just exposing how pants AMD's stock cooler is and that it's spoiling an otherwise solid piece of engineering, if Nvidia and the card partners can source a decent cooler then so can AMD, its very surprising that for the sake of a few $ they don't and run the risk of negative reviews. I would have expected the stock cooler to be high-ish quality, and that the partner boards would instead be just looked to for budget and super-extreme options rather than to have to fix a sub-par reference design.
What happens if we take a Titan or 780 and dial it up so it consumes the same power as a stock 290X, how is the noise level and performance comparison then? Graphics cards need to be compared in several ways, perf/$, per/W and per/dB... different users have different priorities and it's not always about absolute outright frames per second. Personally I'd buy a quieter and more power frugal card even if it was a bit slower and/or cost a bit more.
My point exactly. Making such an accusation based on one sample is unprofessional at best.
As has been mentioned elsewhere, the 290X consumes less power than 780 in as many games as it consumes more, it seems. Also as has been said, heat can have a fairly big impact on static power consumption. I'm repeating myself from another thread, I know, but since heat essentially equals power consumption, a 290X would run at roughly the same temps as a 780 given the same cooler, give or take a few C for heat density variations. I know this isn't the argument you're making but the 'AMD had to run the GPU to its limits to compete' argument we keep seeing is utter nonsense, as having a less-than-ideal cooling solution harms performance if anything.
Oh and of course the 20W or so either way people are squabbling over is frankly ridiculous anyway, as is the argument to spend £100 more on an equally performing GPU purely to 'save power', which will probably be somewhere in the area of £20 per year MAX, if the system was running games literally 24/7.
This one million times over, i don't understand why they are preventing manufacturers from releasing custom cooled cards. Despite what AMD might say about the cooler being sufficient, people want a better cooler!
It's such unbelievably bad PR i cannot fathom it at all, if the custom ones came out they would likely fly off the shelves!
Tarinder said end of November was when AIB models are being shipped.
I have a suspicion AMD knew about the GTX780TI,so I wonder whether the delay is to make sure custom pre-overclocked cards can match it??
Originally Posted by Thracks View Post
I'll chip in with official word on this.
Two outlets in hundreds sampled have uniquely reported instances of AMD Radeon R9 290(X) boards purchased in retail that have exhibited an uncharacteristic level of performance variance as compared to press samples issued by AMD. As retail products purchased by almost every other outlet (e.g. Sweclockers) do not demonstrate this phenomenon, we’re working to secure the aberrant board(s) in question for further analysis. In the meantime, we’ve identified areas where variability can be minimized and are working on a driver update which will minimize this variance.
We will be releasing that driver in the next 24 hours to fully correct this behavior by normalizing fan RPMs vs. PWM control. The 290X in Quiet mode should be at 2200RPM, and the 290 should be at 2650RPM. If anyone is seeing fan speeds below that, they are affected by this issue, and the driver will resolve the issue in 24 hours.
//EDIT: All boards should perform similarly to the AMD-issued samples seen in reviews. Period.
Because they were released with price being a key ingredient. Custom coolers would add £s, and people buy from benchmarks not noise/heat ratings, at least in the early release cycle.
It's not bad PR, since everyone can see what an awesome card it is - it's going to be like there are two release days on these cards; they sold a shed-load on reference release, and will sell a shed-load on custom-cooler release. It's genius.
What is quite funny is that people have very short memories indeed.
Nvidia cards use non-deterministic boost,with a minimum defined range,wheras AMD has always had a defined upper limit.
This lead to problems with reference GTX660TI,GTX760 and Geforce Titan cards,where websites like Hardware Canucks,hardware.fr and pcgameshardware exposed card throttling over extended game runs.
Edit!!
Here are some of the articles:
http://www.hardwarecanucks.com/forum...xy-msi-21.html
http://translate.googleusercontent.c...PcX-0XdORhBOyQ
http://translate.google.com/translat...1056659%2F4%2F
http://translate.google.es/translate...t%2Findex8.php
Last edited by CAT-THE-FIFTH; 06-11-2013 at 06:23 PM.
There are currently 1 users browsing this thread. (0 members and 1 guests)
Posting Permissions
Reply With Quote
