AMD may be cherry picking review samples

[DudeWazap]

NVIDIA does it to, whats your point? that companies never lie? what dream world do you live in?

So we cant live ina world filled with sun shine, rainbows and ponys...=(

[TheMissxu]

I wish they cherry picked parts for me...

Oh wait they did, just they tried to sell it for $830.

 

It's called the silicon lottery, have you heard of it?

[xkronusx]

Lmao are you seriously going to take advice from toms that heavily?

 

They are by far the most bias site i have come across, they have a passionate love for the blue and green team.

 

I would take a look at every site you can find and make an informed decision on whether or not

AMD is playing dirty. I am in no way saying they are not, but do some homework first.

[Lays]

everyone does this, who gives a pickle..

[juretrn]

And this is news how?

[LAwLz]

lol this is the worst article ever and just shows that Tom's Hardware doesn't understand how the card works. 290s are performing the same as 290xs because they have more thermal headroom, and run more consistently at higher boost clocks than the 290x can with the reference cooler. 

But, it also has fewer cores. Yes it has more headroom for overclocking but each MHz increase will give you less performance compared to the 290X. A 10% overclock on a 290X is a much bigger performance increase than a 10% overclock on the 290. Seems strange that both Linus and Tom's hardware both got bad overclocking 290Xs, but golden sample 290s. Might be a coinsident so I wouldn't really accuse AMD of cherry picking review samples (which is a really bad thing and should never be done by anyway, anyone who does it deserves a ton of shit thrown at them).

 

About Intel and Nvidia doing the same, is there any proof? Last time I checked, neither Nvidia nor Intel cherry picks review samples according to reviewers such as Linus (pretty sure he said that in one video, and other reviewers have said so in the past as well).

[SKOONJar]

Hmmm this explains why the 290x that linus got is a poor overclocker

[digitalnav]

We'll find out for sure once we get our supply finally comes in.

[IdeaStormer]

Would be cool to see a comparison of the GPU's manufactures send to be tested by testers and then have the testers also test some that they bought off the street (so to speak), to see if those are the same. Would be very interesting to prove/disprove this binning/cherry picking theory. Results would have a big implication on sales.

[ACatWithThumbs]

Hmmm this explains why the 290x that linus got is a poor overclocker

Because there isn't even one R9-290X that can overclock well, because 28nm is on the limit.

In fact the R9-290X can't even get most time the 1GHZ clock it's normaly between 800-900mhz after 5min of gameplay.

The R9-290X is just to hot for overclocking.

 

[IdeaStormer]

Actually, these benchmarks should be done by a manufacture refunding the price of the part to be tested from the testers buying it at their local reseller (store) first so its a totally random test and not a cherry picked test. Manufacture should not know which reseller (store) it was bought from till after the test. This way the testers can buy from the on-line cheapo store or the highest priced store or both and then get a true sampling.

[mr moose]

which is why you wait for a better average of reviews and benchmarks before you speculate on price/performance or build quality.

[eshbop]

But, it also has fewer cores. Yes it has more headroom for overclocking but each MHz increase will give you less performance compared to the 290X. A 10% overclock on a 290X is a much bigger performance increase than a 10% overclock on the 290. Seems strange that both Linus and Tom's hardware both got bad overclocking 290Xs, but golden sample 290s. Might be a coinsident so I wouldn't really accuse AMD of cherry picking review samples (which is a really bad thing and should never be done by anyway, anyone who does it deserves a ton of shit thrown at them).

 

About Intel and Nvidia doing the same, is there any proof? Last time I checked, neither Nvidia nor Intel cherry picks review samples according to reviewers such as Linus (pretty sure he said that in one video, and other reviewers have said so in the past as well).

At stock 290 runs 12-15% higher clocks than 290x, it has 9% fewer cores. That goes a long way to describing why there is only a few frames between these too cards in these reviews.

 

Reviews of Haswell CPUs have them hitting 4.7-4.8Ghz with relatively low volts (no more than 1.3v), some getting 4.9-5Ghz. TTL was sent 2 Haswell chips that hit 4.9Ghz at less than 1.3v. He also said in his 290x review that Nvidia had sent him (and do send) cherrypicked GPUs and was wondering why it appeared AMD hadn't.

[martinrox1568]

Morals? Isn't that like borderline fraud? What if a keyboard maker would give reviewers an amazing version of a keyboard handmade just for them, then sell a crappy one to the public? It's not that much different.

Amd or nvidia doesn't garuntee over clocks. Hell, some even void the waruntee. Reviewers don't have to over clock and then everything is entirely fair. At stock, even the best overclocker matches the worst. The product still sells as advertised

[ajoy39]

But, it also has fewer cores. Yes it has more headroom for overclocking but each MHz increase will give you less performance compared to the 290X. A 10% overclock on a 290X is a much bigger performance increase than a 10% overclock on the 290. Seems strange that both Linus and Tom's hardware both got bad overclocking 290Xs, but golden sample 290s. Might be a coinsident so I wouldn't really accuse AMD of cherry picking review samples (which is a really bad thing and should never be done by anyway, anyone who does it deserves a ton of shit thrown at them).

 

About Intel and Nvidia doing the same, is there any proof? Last time I checked, neither Nvidia nor Intel cherry picks review samples according to reviewers such as Linus (pretty sure he said that in one video, and other reviewers have said so in the past as well).

The 290xs aren't actually overclocking though, if you go read PcPers review their 290x only actually ran around 800 mhz most of the time once it had heated up, so we're talking about a 200mhz difference. That actually can close the performance gap. 

[Blue fellow]

I'd rather review samples be consistently high, then consistently unknown.

[Equlix]

when you go to a job interview you wear a suite right? even though the job may not require one, same thing here. This is a non-issue blown way out of proportion.