AMD playing dirty tricks with 6850 samples?

Last edited:
At HardOCP we always verify before we evalaute the product that the specifications are correct, notice we always put up the GPUz screenshot on the test page so you know what we are evaluating. In our 6800 Series evaluation, note we did put the GPUz screenshot up, and its showing the correct number of SP units. Just want to make that clear as far as this site is concerned.

The official response from AMD was sent to us today, I wasn't even aware of it until a few hours ago:

Apparently a small number of the AMD Radeon HD 6850 press samples shipped from AIB partners have a higher-than-expected number of stream processors enabled.

This is because some AIBs used early engineering ASICs intended for board validation on their press samples. The use of these ASICs results in the incorrect number of stream processors. If you have an HD 6850 board sample from an AIB, please test using a utility such as GPU-z to determine the number of active stream processors. If that number is greater than 960, please contact us and we will work to have your board replaced with a production-level sample.

All boards available in the market, as well as AMD-supplied media samples, have production-level GPUs with the correct 960 stream processors.
Click to expand...

We will continue to make sure all cards we evaluate are at the correct specs before we evaluate their performance.

This is not a ploy by AMD, rather a mistake some board partners have made. All media samples sent out by AMD for the launch were correct.
 
Reminds me of the thing back when about Nvidia 'stacking' their review cards, a lot of which were sent from EVGA, not Nvidia themselves.

All companies do this kinda crap, if you actually look around.
 
AMD's like that, the competition's like "here's an overclocked card to compare against the competing stock speed card". AMD has bulletholes in their feet from being so nice and honest.
 
sirmonkey1985 said:
completely agree but then the problem lies on the reviewers not doing their jobs and checking the hardware they receive before doing the review..
Click to expand...
A reviewer shouldn't have to babysit hardware manufacturers. Of course, it seems that they have to, but it's the hardware manufacturer that is supposed to be honest.
 
What I am seeing so far is on average the 6850 and the 6870 are really no different than the previous generation 5800 series, sometimes a bit slower BUT a Shitload cheaper. This puts me in the ballpark for a 6970 or two when they release. If 5-10fps saves me $400 when I set out to upgrade this is a good thing. I bought a couple of GTX465's back in August, not to impressed with them so my 4870X2's are back in the saddle until then. :D AMD playin tricks? How about all those Nvidia 9800gtx's resurrected and renamed to fucking hell and back? What video card company still in business isn't doing this?
 
TPU got one with 1120 and so did Rage. Other than them I'm unaware of other sites that did. TPU corrected it so their results wouldn't be skewed and Rage pulled their review.

In any case, the site that counts got it right > [H]ard.
 
I dont know if amd was playing dirty tricks, I know the OP states he jumped the gun with the title, but IMO a engineering sample should NEVER be reviewed. Because as was obvious they are with GPU's that aren't finalized. Those GPu's do not completely resemble the retail copy EVEN if you flash the bios etc..

Nothing short of a card boxed for sale should be reviewed to be ethical and maintain the integrity of the results consumers will experience. Unfortunately everyone rushes to get these reviews done and the end result is the first card they can get is a engineering sample. The end result is a review that should be ignored if seriously considering purchasing this GPU. Good thing there are over 10 sites with valid reviews that dont cast any doubt on what the retail product will offer its lucky customers.
 
Did you guys read what Brent wrote?

This is not a ploy by AMD, rather a mistake some board partners have made. All media samples sent out by AMD for the launch were correct.
Click to expand...

Why are people still speculating as whether this is a "ploy."
 
heflys20 said:
Did you guys read what Brent wrote?



Why are people still speculating as whether this is a "ploy."
Click to expand...

Because the cards themselves are fantastic and destroy Nvidia's only card that was even worth considering. Now people can be drama queens.
 
This reminds me back in the day when motherboard companies would ship their product to be reviewed with the default FSB at 105mhz (or higher) even when it was set to 100mhz in the BIOS.

All the samples that had the wrong shader count were either Sapphire or HIS, right? If you wanna get angry and blame someone, I'd start with those two companies.
 
^ they didn't do it on purpose, its the motivation that is the important factor, and as far as i've seen there was no ill intent, accidents happen, glad it was caught and made public, AMD stood up and admitted the issue, so its not like this is a conspiracy

shoot, i'd be thrilled if i bought a retail card that had more shaders enabled on it than was supposed to be
 
Brent_Justice said:
^ they didn't do it on purpose, its the motivation that is the important factor, and as far as i've seen there was no ill intent, accidents happen, glad it was caught and made public, AMD stood up and admitted the issue, so its not like this is a conspiracy

shoot, i'd be thrilled if i bought a retail card that had more shaders enabled on it than was supposed to be
Click to expand...

If one or both of my Sapphire cards coming this week have more shaders than intended, I'll be tickled pink. :cool:

If this was an accident, I'd be a little worried that their quality control is a little lacking. "Oh you mean the 6850 we were testing/developing with isn't the same as the one we're using for production?" Yikes.
 
Hemi345 said:
If this was an accident, I'd be a little worried that their quality control is a little lacking. "Oh you mean the 6850 we were testing/developing with isn't the same as the one we're using for production?" Yikes.
Click to expand...

More like in the rush to ship things off to reviewers, some intern on his first week picked up the wrong box :p
 
WorldExclusive said:
Who's buying 6850s anyway? Yawn.
Click to expand...

I bought one it is roughly equal to the gtx 460 1gb in performance and price. I needed a new video card so might was well get the newest (also my rig is amd anyways so might as well go with something to compliment it).
 
munkle said:
I bought one it is roughly equal to the gtx 460 1gb in performance and price. I needed a new video card so might was well get the newest (also my rig is amd anyways so might as well go with something to compliment it).
Click to expand...

lol, just kidding, but it's fine if you needed a new card. Just the fact that there's controversy around the 6850 is yawn.
Now if they "purposely" did a switch with the 6970 to have the fastest single gpu, then it would be a story.
6850 still beats the 460 with 960 shaders.
 
I wonder about the Tom's Hardware review. The results look like what you'd expect. They never mention anything aobut a different shader count, but they don't show a gpu-z screenshot so we may never know.
 
If you fellows who are getting a 6850 want to OC them now and not be limited to 850 on the core from the CCC add Force Unsupported GPU = 1 to the config file in MSI's Afterburner folder. You do need to enable unofficial overclocking like normal as well.

It still won't allow you to increase voltage however it will open your core range up.

I'm testing one myself that I will be passing on to my nephew in a week or so when I'm done. I've had it up to 960 so far without voltage increases. That was only for one run though. I haven't had time to thoroughly test it. I can say it does 930|1150 with the fan @ 56% with zero problems. Once proper voltage support arrives I wouldn't be surprised if it hits 1010. This is a Sapphire HD 6850.
 
You must log in or register to reply here.
Back
Top