Retail AMD Fury X Sound Testing - Pump Whine Investigation by PCPerspective

[Takata]

Ironic? I don't think you understand what that is.  PCpers testing actually shows the retail units to be worse than the review sample, they don't "think" it's worse, they "know" it's worse.

 

PCper's article and testing was done such a way as to ascertain the the level of noise coming from the card and if the claims made by AMD where in fact true or not.  Whether the particular noise is attenuated by a case or only displeasurable to certain ears is moot.    It turns out AMD's claims to have already fixed it were somewhat erroneous.

By attempting to fix a problem, they made it worse... is that ironic, or is that the wrong word? Maybe just call it a failure? Nevermind the high frequency offending sound from the review samples. The retail samples drown that out with something a lot louder and very different. On the plus side, there's a lot of spare space inside the Fury X for AMD to experiment with pump redesigns or just stuff it with sound insulation.

 

It is not irrelevant how the case attenuates this noise. If something was loud at a frequency that would be sufficiently attenuated by even a mediocre case, then that's okay. Taking the case away doesn't neccesarily make the test fair, although the conclusion that AMD's modification does not address the original criticism seems to be perfectly fair. I still wish they'd found a mediocre (in terms of sound dampening) case to test the cards in.

 

The review sample sounding a lot better than the retail sample makes me question the use of review samples at all... would be better if reviewer's samples passed through a retailer, rather than going direct from developer to reviewer.

[Majestic]

I think the problem is to do with how 'clean' the power is, and cheaper PSU just aren't as good at lower loads than the more expensive ones (and both of my PSU cost well over $140 each-and I found out later that they are overpriced and used cheap components  )

 

Yeah, i've not experienced a lot of coilwhine from my maxwell cards (3). Used a BeQuiet! Sraight Power 10 400W and a Corsair RM550. 

I mean i can hear them whirring a bit, but only with my noctua fans at 12.5% PWM and late at night.

[Dabombinable]

Yeah, i've not experienced a lot of coilwhine from my maxwell cards (3). Used a BeQuiet! Sraight Power 10 400W and a Corsair RM550. 

I mean i can hear them whirring a bit, but only with my noctua fans at 12.5% PWM and late at night.

Those are better quality than the Corsair VS650 and CX850M.

[mr moose]

By attempting to fix a problem, they made it worse... is that ironic, or is that the wrong word? Maybe just call it a failure? Nevermind the high frequency offending sound from the review samples. The retail samples drown that out with something a lot louder and very different. On the plus side, there's a lot of spare space inside the Fury X for AMD to experiment with pump redesigns or just stuff it with sound insulation.

 

Maybe I read your post backward, anyway, I'd call it embarrassing rather than ironic.

 

 

 

 

It is not irrelevant how the case attenuates this noise. If something was loud at a frequency that would be sufficiently attenuated by even a mediocre case, then that's okay. Taking the case away doesn't neccesarily make the test fair, although the conclusion that AMD's modification does not address the original criticism seems to be perfectly fair. I still wish they'd found a mediocre (in terms of sound dampening) case to test the cards in.

 

 

 

The reason the case is irrelevant is because the article addresses only the claims that the issue had been fixed.  Whether your case or my room makes it a none issue is irrelevant as it has no bearing on certain claims made or the points the article addresses.  The test would not have been fair if they tested them in a case because they need to remove as many variables as possible.  (comparing cards to each other not the environment they are in).

 

 

 

The review sample sounding a lot better than the retail sample makes me question the use of review samples at all... would be better if reviewer's samples passed through a retailer, rather than going direct from developer to reviewer.

 

Review samples are not supposed to be picked, and to be honest we can only trust AMD here that they haven't cherry picked the samples.  I think this is something PCper was trying hard to avoid insinuation on in the article, by pointing out he doesn't think it was a "golden sample".  We are also in the same position with Nvidia, we have no way to know they aren't cherry picking the samples, personally I just choose to believe they are all cherry picked from all companies.

[Sakkura]

Is that store like bad or what ? the one with 980Ti cheaper and on your store I have to pay 100DKK for a 980Ti that's seems like a great price then.

 

The other store I linked to is Computersalg, one of the go-to Danish etailers for computer hardware. It's more niche than Komplett, which is like the Scandinavian Newegg except with much better support.

 

Komplett is usually the first place to have new hardware in stock, but other than that I don't see any reason to prefer them over Computersalg. Nicer-looking website, I suppose.

[Misanthrope]

It's exactly the problem. The 970, before the vram issues, was the go to card for mainstream non-budget builds. A lot of people new to PC building pick a cheap power supply. Even if it's reliable, how clean the power is and how stable it is has a lot of effect on this stuff. Switching from a generic PSU to a decent corsair one fixed my coil whine problem. Does that mean the 970 doens't do as good of a job dealing with "dirty" power? Idk

 

As I understood this this isn't coil whine, just pump noise. So it has nothing to do with electricity and power delivery and it's pretty much a mechanical sound, much like a bad noisy fan only thing to do is replace it.

 

Which is shitty because of the custom vram cooling that doesn't seems to have any kind of standard fittings, otherwise you'd be able to swap it for an AIO or custom block and call it a day

[Sakkura]

Why are reference coolers never the open type like Gigabyte's Windforce or Sapphire's Vapor-X (on the 290X) anyway?

 

AMD could just use that all that empty space inside the Fury X for sound dampening material.

 

AMD actually tried that with the Radeon HD 7990. It didn't work out all that well, although it was much better than the blower-style 6990. The R9 295X2 was a huge improvement though.

[Takata]

Review samples are not supposed to be picked, and to be honest we can only trust AMD here that they haven't cherry picked the samples.  I think this is something PCper was trying hard to avoid insinuation on in the article, by pointing out he doesn't think it was a "golden sample".  We are also in the same position with Nvidia, we have no way to know they aren't cherry picking the samples, personally I just choose to believe they are all cherry picked from all companies.

Linus' sample was dead (or at least artifacting) on arrival. Not sure how to interpret that. ?

 

 

Which is shitty because of the custom vram cooling that doesn't seems to have any kind of standard fittings, otherwise you'd be able to swap it for an AIO or custom block and call it a day

You _could_ get the waterblock for the Fury X from EK, but building your own loop would add hundreds of dollars to the cost.

[mr moose]

Linus' sample was dead (or at least artifacting) on arrival. Not sure how to interpret that. ?

 

 

 

 

I wouldn't read into it,  sometimes shit just happens, it's the first time I have heard of a review sample being dodgy, but law of averages says it has to happen at least once.

[Rohith_Kumar_Sp]

I wouldn't read into it,  sometimes shit just happens, it's the first time I have heard of a review sample being dodgy, but law of averages says it has to happen at least once.

true and true - i've bought 2 nvidia cards that were dead with artifacts and had to return it the next day, first it was gt9400, all i could see was pixelated mess, the 2nd was a 780ti, bought it, had it installed on the build on spot, had artifacts all over, replaced it on spot, and the retailer was pissed as he knew before everyone on the internet that next week 900 series was gonna launch because of the distributer, and i wanted an MSI he made a sales pitch for Asus to get rid of the cards, i got an Asus but the card was DOA, and he got pissed that he lost money on it.

sometimes, stuff happens for no reason. doesnt mean nvidia cards were bad as in case of Linus all Fury card were bad or defect

[Misanthrope]

Linus' sample was dead (or at least artifacting) on arrival. Not sure how to interpret that. ?

 

 

You _could_ get the waterblock for the Fury X from EK, but building your own loop would add hundreds of dollars to the cost.

 

Also it might not make sense after the Fury nonx is released: it will be cheaper and I bet there will be custom blocks for it and we don't know performance yet but I'm willing to bet you'll be able to overclock to fury x levels, see 290 vs 290x

[cesrai]

Also it might not make sense after the Fury nonx is released: it will be cheaper and I bet there will be custom blocks for it and we don't know performance yet but I'm willing to bet you'll be able to overclock to fury x levels, see 290 vs 290x

EK made a custom block for Fury X btw.

[Trik'Stari]

Seems to me like AMD should have allowed the other manufacturers to make their own fury X cards. Perhaps this wouldn't have happened.

 

But then again, who knows. AMD just needs to get it together.

[Misanthrope]

EK made a custom block for Fury X btw.

 

Yes I saw it and it's damn nice looking btw. Still a bit hard to justify on an already watercooled card but man does it looks good

[Briggsy]

Cooler Master probably just repurposed old 120V pumps. I owned one (for about 10 minutes before tossing it) and the 120V pump whine vs. FuryX whine is very similar, in that it feels like its piercing your skull.

 

[cesrai]

Cooler Master probably just repurposed old 120V pumps. I owned one (for about 10 minutes before tossing it) and the 120V pump whine vs. FuryX whine is very similar, in that it feels like its piercing your skull.

 

That sounds interesting.

[LAwLz]

The 980ti is 200$ more expensive than the Fury X in Denmark, and the Fury X comes with an AIO included in the price. Vastly superior temps on the GPU.

 

But yeah, it sucks that the pump makes that noise. No one likes that.

Sigh... Can you never be honest about anything AMD related?

Here are the 980 Ti prices in DK (click on one card to see what it costs in multiple stores) and here is what the Fury X costs in DK.

 

The Fury X is 6659 DKK (990 USD)

 

The 690 Ti is 6600 DKK (980 USD) for the cheapest reference card, and it's in multiple stores such as CDON which are very trusted. The 980 Ti reference design is cheaper than the Fury X, not 200 dollars more expensive.

 

The 980 Ti G1 costs 7200 DKK (1070 USD) from Komplett (same store which sells the cheapest Fury X) so your "it's 200 dollars more expensive" claim doesn't even hold up when comparing the cheapest Fury X vs the best 980 Ti aftermarket design. That comparison only ends up being an 80 dollar difference.

 

 

Edit: Never mind I am an idiot. Mistook Norway for Denmark.

[MEC-777]

This just in today: http://www.techpowerup.com/213993/amd-revises-pump-block-design-for-radeon-r9-fury-x.html

 

Apparently they have taken action to fix the pump whine. But of course, there will still be those from the "first run" out on the shelves and people will still be receiving them from retailers. If I were considering buying one, I'd probably wait a few weeks to a month or so to allow the new stock with the revised pumps to start shipping. 

 

PCper's investigation might have been a bit pre-mature. Fixing something like this doesn't happen overnight. It can take weeks or even longer. But, it looks as though they may have actually fixed it.

 

 

(@BiG StroOnZ - This should probably be added to the OP as an update so people don't continue on thinking it hasn't been fixed and that AMD has done nothing. There are enough false claims about AMD spouted on a daily basis as it is. ) 

[Rohith_Kumar_Sp]

This just in today: http://www.techpowerup.com/213993/amd-revises-pump-block-design-for-radeon-r9-fury-x.html

 

Apparently they have taken action to fix the pump whine. But of course, there will still be those from the "first run" out on the shelves and people will still be receiving them from retailers. If I were considering buying one, I'd probably wait a few weeks to a month or so to allow the new stock with the revised pumps to start shipping. 

 

PCper's investigation might have been a bit pre-mature. Fixing something like this doesn't happen overnight. It can take weeks or even longer. But, it looks as though they may have actually fixed it.

 

 

(@BiG StroOnZ - This should probably be added to the OP as an update so people don't continue on thinking it hasn't been fixed and that AMD has done nothing. There are enough false claims about AMD spouted on a daily basis as it is. ) 

i mean where was this in QA? they're just losing money over replacements 

[BiG StroOnZ]

This just in today: http://www.techpowerup.com/213993/amd-revises-pump-block-design-for-radeon-r9-fury-x.html

Apparently they have taken action to fix the pump whine. But of course, there will still be those from the "first run" out on the shelves and people will still be receiving them from retailers. If I were considering buying one, I'd probably wait a few weeks to a month or so to allow the new stock with the revised pumps to start shipping.

PCper's investigation might have been a bit pre-mature. Fixing something like this doesn't happen overnight. It can take weeks or even longer. But, it looks as though they may have actually fixed it.

(@BiG StroOnZ - This should probably be added to the OP as an update so people don't continue on thinking it hasn't been fixed and that AMD has done nothing. There are enough false claims about AMD spouted on a daily basis as it is. )

LOL Their source is a single forum post, claiming it has been fixed.

[MEC-777]

i mean where was this in QA? they're just losing money over replacements 

 

They probably released the first batch just to get some samples/product out there, or it would have been delayed a week or so longer for release (which would have been a better idea, IMO, instead of releasing noisy cards but I digress. ).

 

LOL Their source is a single forum post, claiming it has been fixed.

 

That's why I said "apparently" and "may" have been fixed.

 

AMD stated that they knew of the problem and that it would be fixed in retail versions, but they didn't say "all retail versions from day one". I think they should have waited to ship the first run of cards and fixed the problem before any retail versions hit the shelves/customers/reviewers. But, that would have pushed back availability and they've already kept us waiting longer than they should, so it's a bit of a trade off I guess.  

[Sakkura]

Sigh... Can you never be honest about anything AMD related?

Here are the 980 Ti prices in DK (click on one card to see what it costs in multiple stores) and here is what the Fury X costs in DK.

 

The Fury X is 6659 DKK (990 USD)

 

The 690 Ti is 6600 DKK (980 USD) for the cheapest reference card, and it's in multiple stores such as CDON which are very trusted. The 980 Ti reference design is cheaper than the Fury X, not 200 dollars more expensive.

 

The 980 Ti G1 costs 7200 DKK (1070 USD) from Komplett (same store which sells the cheapest Fury X) so your "it's 200 dollars more expensive" claim doesn't even hold up when comparing the cheapest Fury X vs the best 980 Ti aftermarket design. That comparison only ends up being an 80 dollar difference.

 

Uh... all of that is from Norway, not Denmark. Those prices are in NOK, not DKK. For Denmark, I would recommend EDBpriser.dk over Prisjakt.no.

 

The Fury X is available for 5519 DKK at Komplett, while in an earlier post I linked a GTX 980 Ti from Computersalg for 5550 DKK.

[LAwLz]

Uh... all of that is from Norway, not Denmark. Those prices are in NOK, not DKK. For Denmark, I would recommend EDBpriser.dk over Prisjakt.no.

 

The Fury X is available for 5519 DKK at Komplett, while in an earlier post I linked a GTX 980 Ti from Computersalg for 5550 DKK.

Well crap... Pricerunner has a danish site too apparently so I'll use that instead.

 

980 Ti vs Fury X.

 

Fury X - 5900 DKK (878 USD)

 

980 Ti Reference - 5600 DKK (834 USD) (plus/minus 100) from multiple stores such as ProShop, ComputerSalg and CDON.

 

980 Ti non-Reference - 6300 DKK (938 USD). Two different models, MSI Gaming or Zotac AMP.

 

So I might have mistakenly looked at Norwegian prices, but my point still stands. There is not a 200 dollar price gap in Denmark, even if you are looking at non-reference cards. If you're looking at the reference 980 Ti then it's actually cheaper.

 

Stop lying to make AMD seem better/Nvidia seem worse than they are. Please.

[R_Golden]

That comments section is aids, so many fanboys.

[Notional]

Sigh... Can you never be honest about anything AMD related?

Here are the 980 Ti prices in DK (click on one card to see what it costs in multiple stores) and here is what the Fury X costs in DK.

 

The Fury X is 6659 DKK (990 USD)

 

The 690 Ti is 6600 DKK (980 USD) for the cheapest reference card, and it's in multiple stores such as CDON which are very trusted. The 980 Ti reference design is cheaper than the Fury X, not 200 dollars more expensive.

 

The 980 Ti G1 costs 7200 DKK (1070 USD) from Komplett (same store which sells the cheapest Fury X) so your "it's 200 dollars more expensive" claim doesn't even hold up when comparing the cheapest Fury X vs the best 980 Ti aftermarket design. That comparison only ends up being an 80 dollar difference.

 

 

Edit: Never mind I am an idiot. Mistook Norway for Denmark.

 

That's the problem with having 2 threads. I already went through updated prices and availability in the master thread yesterday. But it is correct that you can now get a 980ti for cheaper. But as I pointed out in the other thread, supply and demand, and entry to market plays a big role on pricing. So the Fury X is priced highly, as there are few of them. The cheapest store for the 980ti for instance has price gouged the hell out of the Fury X, at over 10k DKK. Also HBM and AIO's are value adding propositions in themselves.

 

Reference versus reference is always most fair, but if you factor in cost and functionality, maybe EVGA's AIO version is more fair, as AIO cost drives price up. Either way, conclusion is that 980ti is cheaper right now in DK. Have you thought about why NVidia, usually having the price premium, don't anymore? Competition is a great thing

 

As for on topic,; PCPer, did state that their 2 retail versions are the Cooler Master sticker versions, and that they knew about the Cooler Master embossed versions. They also said they would buy two more retail versions with embossed CM units and test those.