AMD accused of "Golden Sample" on 290X given to reviewers, retail bought cards throttling

[straxusii]

I always thought this was a given for all hardware, not always but I guess when given the choice of which part to give the highest profile reviewer of course you give them the one that clocks the highest.

I doubt any special engineering is going on though.

 

[Gibbo]

I always thought this was a given for all hardware, not always but I guess when given the choice of which part to give the highest profile reviewer of course you give them the one that clocks the highest.

I doubt any special engineering is going on though.

All I can say is we had two press samples.

One did 1160 core region, the other 1175-1200, both with Asus BIOS flashed. Memory wise they both did 6400-6600MHz.

Sapphire (Asus BIOS) R290X did 1200 core easy and 6400MHz RAM.
Asus R290X did 1200-1220 core and 6600-6800MHz RAM.

Both better than press samples.
None of our cards experienced throttling in testing even with all cards in quiet mode at stock.

 

[Rasch]

OFC it throttles if it reaches the target temp.If i use it at 40% fan it throttles dowm to keep the temps at 94C,but at 55% it never throttles.Changing the cooler and you got a beast which u can overclock decently and absolutely annihilates anything.Just check out Kaapstad's Firestrike extreme score on AIR,which is 6k!

 

[w3bbo]

I have reviewed GPU's for the past 7 years now and not once can I categorically state that a card direct from AMD or NVIDIA was any faster than one from retail or one of the main re-sellers. Some cards perform better than others but then the same can be said about retail cards. Every card will not clock exactly the same, I have had good and not so good clockers both from AMD/NVIDIA. Like retail, it is luck of the draw or perhaps, just maybe, user error?

Never heard so much tosh in my life.

 

[mmj_uk]

I always thought this was a given for all hardware, not always but I guess when given the choice of which part to give the highest profile reviewer of course you give them the one that clocks the highest.

It is but it's no longer a case of sending the 'highest clocking' ones to reviewers and giving everyone else ones that simply runs at spec, what AMD are accused of is sending reviewers one which runs at spec and everyone else ones which run below it due to early throttling (unless set to uber mode/custom fan profile).

 

[BubbySoup]

So because 2 reviewers got duds at retail and one site got lucky, that means there's nothing to it?
Even of not golden samples, at the very least people are having issues with throttling, which people should want to know about for RMA purposes, no?

AMD have said the retail copy that toms have is faulty, hence the issue, so we should stick our heads in the sand and not make people aware about a potential for getting a faulty card and what to do about it?

So based on your massive sample of 3, you think there is something to it?

 

[Marine-RX179]

All I can say is we had two press samples.

One did 1160 core region, the other 1175-1200, both with Asus BIOS flashed. Memory wise they both did 6400-6600MHz.

Sapphire (Asus BIOS) R290X did 1200 core easy and 6400MHz RAM.
Asus R290X did 1200-1220 core and 6600-6800MHz RAM.

Both better than press samples.
None of our cards experienced throttling in testing even with all cards in quiet mode at stock.

IMO this need to be sticky.

Regardless of the truth, the usual bunch would be spamming this in every topic posting this like how they keep making out the AMD reference cooler noisier than it really is (not saying it is not noisy, but it's been OTT blown out of proportion and being spamming into almost every single thread).

 

[drunkenmaster]

In his 290X review (using a pre-release sample) TTL said that the Nvidia cards he gets sent to him are almost always golden samples that do better than most retail cards.

Wanna make a thread on that?

The difference is Nvidia have a history of going around pushing people who have a financial stake(advertising on sites for example) into publicly trashing AMD whenever possible, AMD don't do this.

So when Nvidia sends out golden samples unmatched by retail cards, reviewers with nvidia or AMD advertising on their site say nothing. When AMD do it(if they did it) then sites who get money from Nvidia get really vocal about it........ imagine my surprise......

 

[andybird123]

All I can say is we had two press samples.

One did 1160 core region, the other 1175-1200, both with Asus BIOS flashed. Memory wise they both did 6400-6600MHz.

Sapphire (Asus BIOS) R290X did 1200 core easy and 6400MHz RAM.
Asus R290X did 1200-1220 core and 6600-6800MHz RAM.

Both better than press samples.
None of our cards experienced throttling in testing even with all cards in quiet mode at stock.

This isn't about overclocking, obviously results will vary when they are pushed... the reviewers are saying that the samples on stock settings attained 1ghz and stuck there on stock settings, take it out of the box and it runs all day at 1ghz... retail bought, same test, throttles to 700mhz

Note to all; *I'm* not accusing AMD of anything... 2 reviewers are, and they are asking for consumers help to do a wider test to ascertain what the chances of getting a problem card are

So based on your massive sample of 3, you think there is something to it?

Not my sample, just relaying information as it was given to me by the media, people are free to infer their own conclusions

 

[ALXAndy]

Because Intel didn't do the same thing with Haswell......

 

[mmj_uk]

Another article here:
http://techreport.com/news/25609/up...9-290x-cards-may-be-slower-than-press-samples

AMD's response:

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. 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. 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 provide an update shortly.

So they don't deny it but can they really fix it via drivers? only way I'd imagine they can do it if it's due to variance in manufacturing is more aggressive fan profiles.

 

[Gregster]

The level of fanboyism in this thread is quite absurd. Some people need to grow up/chill out/enjoy life.

Bad press is always there, regardless if you are a fan of red or green or neither. Don't take things to a personal level.

 

[teppic]

The second site (techreport) is just quoting the first site (toms), which itself bases this 'revelation' on a single press sample versus two retail samples.

 

[andybird123]

Well said.

Andy forgot this part as well.



If anything Toms probably had a faulty card.

Which I already posted
Nice try matt

 

[MDPlatts]

I believe that was the R9 290 which has a default fan speed of 40% which mean't that the card throttled when being tested meaning performance wasn't what it could be.

AMD released a new beta driver which was set to override the bios setting and change the fan speed to 47% which gave the 290 much better performance.

I imagine AMD will change the bios setting on new cards produced set at 47% for the fan speed.

The SPEED-LIMIT is 40% on the 290X and 47% on the 290 - not the SPEED. If its cool then the fan runs slower/quieter/silently and ramps up when the chip gets hotter as required, and when it hits the fan limit, temps rise until they hit 94c, then clock speeds are throttled to keep below that temp.

As ever, one persons case/cooling/airflow/temps will be different than anothers - and if the case has good/cool airflow in from below then you don't even get close to these limits except when benching but if you put it in shoebox with little airflow then it (and probably any other card) will be noisy.

As for the story - utter rubbish as gibbo has said.

 

[spixelspixel]

This isn't about overclocking, obviously results will vary when they are pushed... the reviewers are saying that the samples on stock settings attained 1ghz and stuck there on stock settings, take it out of the box and it runs all day at 1ghz... retail bought, same test, throttles to 700mhz

No reviewers say the samples stayed a 1ghz, they all showed throttling once it reached the temp limit, just check the original tomshardware review. According to tomshardware, the retail one dropped to whatever speed then stayed there while the sample core clock varied giving it much better performance.

 

[Suarez7]

I can imagine both nvidia and AMD test the cards they send out for review but I don't imagine they test max overclocks, just whether the card works as it should, no throttling, crashing etc etc.

 

[Uncle Petey]

We have a fair number of 290X owners on the forum now. Perhaps they'd care to run a few tests to see whether their cards throttle down to 700mhz?

As has been mentioned in one of those links and by people here, the problem is the 40% fan speed.
The Tom's guy seems to have a slightly dodgy card if you ask me (and AMD).

 

[Gregster]

Update: As is Tom's Hardware policy, we shared these potentially problematic findings with AMD prior to publication, and the company insists something is wrong with the retail-purchased cards I tested. We will continue investigating and, if any additional news becomes available, update this story.

So looks like Toms had faulty cards if they were throttling according to AMD

 

[Marine-RX179]

The point is, you'd would imagine reviewer would do the sensible thing of at least testing a couple of more 290x before making (or hinting) such a claim.

They have test ONE retail 290x and is already getting vocal about it...it does sound a bit illogical and unprofessional. By that logic, if a reviewer bought a retail card off the shelf and end up DOA, I guess they should get vocal saying all the cards off the shelf would DOA by that logic?