AMD denies cherry-picking R9 290, 290X review cards

AMD has denied accusations that Radeon R9 290 and 290X graphics boards supplied to publications for review are hand-picked 'golden sample' boards that perform considerably better than the same boards purchased at retail.

AMD's latest R9 family of graphics processing units (GPUs) have caused something of a stir, offering considerable performance at a reasonable price-point with the flip-side of extreme heat exacerbated by a less-than-optimal stock cooler. One site, however, claims that the high performance seen on the AMD-provided review sample isn't reflected when running the same benchmarks on independently-purchased retail hardware - accusing the company of cherry-picking so-called 'golden sample' hardware for review.

In its analysis of the theoretically lower-end Radeon R9 290 board, Tom's Hardware noticed a discrepancy: the AMD-provided R9 290 review sample was often outperforming a more expensive and supposedly more powerful R9 290X board purchased from retailer Newegg.

The site places the blame on the dynamic clocking mechanism, which aims to keep the core temperature of the board below a certain level. This mechanism was seen to drop the retail R9 290X from its theoretical 1,000MHz core clock maximum to just 727MHz - crippling performance. 'For R9 290X, we can go out and buy boards to compare [performance],' the site explained. 'But there’s no way to know if the R9 290s you buy will operate at the top of their range (947 MHz) or the bottom (662 MHz).'

Responding to press enquiries, an AMD spokesperson denied that the company has provided cherry-picked hardware to reviewers. 'A media outlet has uniquely reported instances of AMD Radeon R9 290X boards purchased in retail that have exhibited an uncharacteristic level of performance variance as compared to press samples issued by AMD,' the company's statement read. 'We’re working to secure the board(s) in question for further analysis. Boards purchased by other media outlets have not exhibited similar characteristics that we’re aware of.'

Although it denies any wrongdoing, AMD has admitted that performance variance is an issue with the R9 series. 'We’ve identified areas where [performance] variability can be minimised,' the spokesperson added, 'and are working on a driver update which will minimize this variance.' No release date for the update has been provided.

UPDATE 20131108:
True to its word, AMD has released a beta driver which it claims resolves the issue. The problem, the company has claimed, is in an unexpected variability in fan speeds which impacts the performance of the stock cooler - resulting in certain board batches reaching their target temperatures earlier than others, and performing accordingly poorer.

'We’ve identified that there’s variability in fan speeds across AMD R9 290 series boards,' an AMD spokesperson explained via email. 'This variability in fan speed translates into variability of the cooling capacity of the fan-sink. The flexibility of AMD PowerTune technology enables us to correct this variability in a driver update. This update will normalize the fan RPMs to the correct values. The correct target RPM values are 2200RPM for the AMD Radeon R9 290X ‘Quiet mode’, and 2650RPM for the R9 290.'

The updated driver, currently only a beta, uses the company's PowerTune platform to adjust the fan speeds to match between batches and should, in theory, mean all 290X and 290 boards have the same performance.