AMD denies sending ‘golden sample’ R290 cards to reviewers

T

terror_adagio

n00b
Joined
Nov 3, 2013
Messages
45
Tech site Tom’s Hardware recently questioned the variability of AMD Radeon R9 290(X) video cards following tests of both a review sample supplied by AMD and a shop bought example bought online. In summary, the confused reviewing team found their R9 290 press board to be “typically faster” than an R9 290X bought in retail. Today Bit-Tech has heard from an AMD spokesperson in a statement denying cherry-picking, or providing ‘golden sample’ cards to reviewers but admitting variability issues.
Click to expand...

http://hexus.net/tech/news/graphics/62093-amd-denies-sending-golden-sample-r290-cards-reviewers/
 
I think both AMD and nVidia do this. I've always felt reviewers get the better cards. Its bad practice.
 
Sounds like dell monitors what reviews got and what was sold were 2 different things.

So a review 490 === is, might, probably, almost sure, a 490x.

variability = meaning we got the wrong thing in the box...wonder how that happened
lying means with a different label on the unit...whoops
 
What's the point? Sure, with the new AMD architecture there is a point, since its speed is based on how cool the card runs, but in previous generations? Cherry-picking would only help overclocking results/reviews.

Anyway, I think it more likely that this is pure coincidence, or a simple misunderstanding of how the new architecture works. There *will* be speed differences between these cards, since some cards run cooler than others and thus gain more performance with the new architecture. It's perfectly conceivable that they randomly received a card that happens to be one of the better ones, and it's also conceivable that they happened to buy a card that is one of the worse ones. Either way, there's nothing suspicious about this one case.

If on the other hand this is something that has been happening with all or most reviewers, then we can start worrying.
 
I've seen a fair bit of temperature differences when it comes to identical stock cards, so it could end up being rather large. Maybe up to 10-15% between the best performer and the worst performer? Room temperature will also play a huge part in this. A difference of +- 5-10 degrees Celsius in the room could have a huge impact on performance.

That said, all of that could be fixed by manually setting the fan speed to a non-throttling level, or new drivers from AMD which sets the fan speed ceiling higher automatically. Or simply aftermarket cooling. Overclocking aside, all of the cards should have the same stock potential with unrestricted fan speed.
 
my card stays rock solid at 1050/1350. after a few hours of BF4 Afterburner shows perfectly flat lines for core and memory clock. sounds like Tom's got an anomaly card. well that or i got an anomaly.
 
Porter_ said:
my card stays rock solid at 1050/1350. after a few hours of BF4 Afterburner shows perfectly flat lines for core and memory clock. sounds like Tom's got an anomaly card.
Click to expand...

That, or Toms is nVidia biased or wants to get a lot more hits on their article by creating a little controversy.

I bet both.
 
simple solution put the cards under water.

Amd's Reference cooling solution needs to be improved. I thought they were going in the right direction with the 7990 reference coolers. Guess they decided to keep the blower around.
 
Hahaha this is awesome.

AMD lied about (and got CAUGHT) with frame pacing issues last year. They denied it. Ryan at PCPER.com proved it. AMD backtracked.

Now AMD gets caught sending golden cards to reviewers. Why trust them again?
 
Sprayingmango said:
Hahaha this is awesome.

AMD lied about (and got CAUGHT) with frame pacing issues last year. They denied it. Ryan at PCPER.com proved it. AMD backtracked.

Now AMD gets caught sending golden cards to reviewers. Why trust them again?
Click to expand...

I am going to guess you haven't noticed that Tom's is the only place saying this?

Tom's. That place. H says otherwise a few posts up.
 
ebduncan said:
simple solution put the cards under water.
Click to expand...

I keep seeing people say this, and perhaps on a site like [H] this might be more common. However, it is not a "simple" solution if you don't already have a loop.
 
Sprayingmango said:
Hahaha this is awesome.

AMD lied about (and got CAUGHT) with frame pacing issues last year. They denied it. Ryan at PCPER.com proved it. AMD backtracked.

Now AMD gets caught sending golden cards to reviewers. Why trust them again?
Click to expand...

Nothing has been proven yet. We should ALL(Hint hint) be objective about this until the facts are made clear.
 
I watched a oc review earlier today, and the very overclockable 290 review sample was surpassing the barely overclockble 290x retail in uber mode.

odd i thought.
 
Last edited:
Ocean said:
I watched a oc review earlier today, and the very overclockable 290 review sample was surpassing the barely overclockble 290x retail.

odd i thought.
Click to expand...

when people review the 290 and oc it and compare it to a 290x in quiet mode and say wow
the 290 beats the 290x.

In quiet the 290x throttles.
290x uber mode + 100% fan people get it to 1065 mhz...
 
I dont see a problem with this, why wouldn't they try to make their card look good to everyone? This is a non issue for me.
 
Toms hardware hasn't even been relevant since the Athlon Slot A days... You goto his site and riddled with the most annoying ads.

Tom's has been known to take money and become very partial to that particular brand in the past . Screw their site and their reviews.
 
Nauseous said:
I dont see a problem with this, why wouldn't they try to make their card look good to everyone? This is a non issue for me.
Click to expand...
Well it's an issue if you base your purchasing decision on a review that shows performance that your store bought card won't or can't achieve. That would be dishonest of the manufacturers, misleading and you should care, were this true.
 
I read a interesting article on this and someone touched on this that matches my findings and how I feel about this

Both Nvidia and Amd are noticing a discrepancy in performance because their cards have performance profiles that take temperature into account and throttle the clocks based on that.

What many may be noticing is a bad thermal paste application or heatsinks that aren't properly applied. YMMV on these kinds of things, hence why I always make sure to apply my own thermal compound and test to make sure the heatsinks are properly applied. Hell I've even lapped some heatsinks to gain major performance boosts in the past as the heatsinks are not properly factory flattened and cut.
 
maybe NVidia and AMD should start sending faulty cards from now on, right fellas?
 
ebduncan said:
simple solution put the cards under water.

Amd's Reference cooling solution needs to be improved. I thought they were going in the right direction with the 7990 reference coolers. Guess they decided to keep the blower around.
Click to expand...

Once you go to watercooling, no air cooling solution is "good enough".
 
Did anyone ever check to see if Tom's is comparing the 290 and 290X with the same drivers? In his review he had them on different drivers and was commenting about the performance similarity. Why didn't they just retest both on same drivers?
 
I am pretty sure it is down to heat causing the problems. People with poor airflow in their cases = high temps = lower GPU clock speeds = slower performance. Thats the problem.
 
Every single company that voluntarily sends out their product to be reviewed makes sure that it is cherry picked the best, I would be surprised if they otherwise. Why risk sending an under-performing, or worse yet, faulty product to receive a worse score than they could achieve otherwise?
 
exlink said:
Every single company that voluntarily sends out their product to be reviewed makes sure that it is cherry picked the best, I would be surprised if they otherwise. Why risk sending an under-performing, or worse yet, faulty product to receive a worse score than they could achieve otherwise?
Click to expand...

Because it will be thrown right back at you... did you read the thread?

AMD doesn't cherry pick cards sent to reviewers other than doing extra QC to make sure it isn't a faulty card.
 
exlink said:
Every single company that voluntarily sends out their product to be reviewed makes sure that it is cherry picked the best, I would be surprised if they otherwise. Why risk sending an under-performing, or worse yet, faulty product to receive a worse score than they could achieve otherwise?
Click to expand...

If this is true, why did AMD send an overheating (you can say faulty - because it had to be fixed by a bios update) MSI card to techpowerup?

http://www.techpowerup.com/reviews/MSI/R9_280X_Gaming/31.html

And there must be more examples than this.
 
This has been a pretty bad launch for AMD so far. Between the delays, underwhelming performane, heat/noise/power draw, etc.

They need to get their stuff together and start polishing things better.
 
LordEC911 said:
Because it will be thrown right back at you... did you read the thread?

AMD doesn't cherry pick cards sent to reviewers other than doing extra QC to make sure it isn't a faulty card.
Click to expand...

Did I say that they were sending out cards that performed better than the others? No.

I clearly said exactly what you said as in companies don't send out under-performing or faulty cards; and performing extra QC tests, in some eyes, is considered cherry picking.

Now get your panties out of your ass and calm down.
 
smaudioz said:
If this is true, why did AMD send an overheating (you can say faulty - because it had to be fixed by a bios update) MSI card to techpowerup?

http://www.techpowerup.com/reviews/MSI/R9_280X_Gaming/31.html

And there must be more examples than this.
Click to expand...

Quality control only goes so far, you can test a product for hours, days, weeks and months...but you have no idea how it will perform the second it leaves the facility and is in someone else's hands.
 
You must log in or register to reply here.
Back
Top