AMD's R9 Fury X - Pump Whine Fixed

[mr moose]

The highest number on the scale is based on what the highest frame rate on the chart is. What are you even talking about?

 

 

The charts where the fury x winds is 5 fps per line, the one where it looses is 10. They are all under 50. Plus everything I've seen thus far has those wrong, even though the fury  x does trade blows, not with those games really.

 

 

Second chart spelled TressFX wrong. TessFX what is that? I mean come on you had one job.

 

 

Why did the OP even put the charts in, the issue has nothing to do with the cards performance.

 

And also I agree, having winning charts at one scale and the loosing chart in another is biased and dodgy and there is no excuse for that, I would ignore any reviewer who does that without a really (and I mean really) good reason.

[Takata]

so i went and found where in the pcper podcast the fury x stuff starts and the quote from amd is

starts @ 46:00  and the quote come sin right around 47:40

 

Yep, that definitely claims that is was only samples and not retail product that was effected.  

Eh... I'm not convinced. The quote is:

"The issue is limited to a very small batch of initial production samples and we have worked with the manufacturer to improve the acoustic profile of the pump. This problem has been resolved and a fix added to production parts and is not an issue. "

 

I would say, "initial production samples" does not equal "review samples", unless you are insinuating that there's one production run for review samples, and then a different production run for the first batch to be sold to customers. I think "initial production samples" refers to the ones with stickers, some of which were sold, and some of which were given to reviewers. The "fix added to production parts" probably refers to the embossed pumps, which supposedly have less whine, but we can't be _really_ sure until PCPer puts them in front of their microphone.

 

I'm guessing PCPer can send their cards back and ask for replacement cards with embossed pumps, since AMD considered the initial stickered pump to be an actual problem.

[ZestyJalapeno]

Plus everything I've seen thus far has those wrong, even though the fury  x does trade blows, not with those games really.

The TR and GTA results are very consistent with a lot of the reviews out there. At 4K the Fury X can and does beat the reference 980 Ti in TR albeit only just and of course loses to it in other games. TW3 results can be considered neck and neck on the average and the train wreck that is AC:U shouldn't be even considered a benchmark.

[BiG StroOnZ]

 

I'm guessing PCPer can send their cards back and ask for replacement cards with embossed pumps, since AMD considered the initial stickered pump to be an actual problem.

 

They did, AMD ignored all their requests:

 

Watch Here

 

or: 

 

 

skip to 55:37

[mr moose]

Eh... I'm not convinced. The quote is:

"The issue is limited to a very small batch of initial production samples and we have worked with the manufacturer to improve the acoustic profile of the pump. This problem has been resolved and a fix added to production parts and is not an issue. "

 

I would say, "initial production samples" does not equal "review samples", unless you are insinuating that there's one production run for review samples, and then a different production run for the first batch to be sold to customers. I think "initial production samples" refers to the ones with stickers, some of which were sold, and some of which were given to reviewers. The "fix added to production parts" probably refers to the embossed pumps, which supposedly have less whine, but we can't be _really_ sure until PCPer puts them in front of their microphone.

 

I'm guessing PCPer can send their cards back and ask for replacement cards with embossed pumps, since AMD considered the initial stickered pump to be an actual problem.

 

There is really only one standard definition of "sample" and in this context it does not apply to a run of finished products ready for shipping. you could use semantics to bend this anyway you want, but basically they are either saying they knew it was present in the samples they took from the production run and sold them anyway, or that they are trying to insinuate that the retail versions did not have issues because it was limited to a small number of initial production samples and fixed. Either way they said it was fixed and not an issue,  yet clearly still is for some.

[Takata]

There is really only one standard definition of "sample" and in this context it does not apply to a run of finished products ready for shipping. you could use semantics to bend this anyway you want, but basically they are either saying they knew it was present in the samples they took from the production run and sold them anyway, or that they are trying to insinuate that the retail versions did not have issues because it was limited to a small number of initial production samples and fixed. Either way they said it was fixed and not an issue,  yet clearly still is for some.

I would say that "sample" refers to any subset of a population. In this case the "subset" is the stickered Fury Xs, and the "population" is all the Fury Xs that exist and will exist, both stickered and embossed versions. However I think AMD is using language to minimise the significance of the problem, by saying it's a "very small batch", and that the "problem has been resolved", when in reality, most of the Fury Xs that exist today will have the defective parts, and the problem is not neccesarily resolved for anyone who already owns one or buys one in the near future. AMDs statement walks the line like an experienced spin doctor, not a lie, but easy to interpret in a way that flatters AMD too much.

 

I'm guessing they chose to sell the Fury Xs with defective pumps with the belief that it was more important to get the product out the door with a minor flaw than to give the 980Ti another few weeks without competition. Fixing product flaws or updating incomplete products later is a strategy that works for almost anything.

 

 

They did, AMD ignored all their requests:

 

-snip-

It sounds like he was asking for information and samples as a reviewer, as opposed to asking for a replacement as a customer with a defective product, which I'm guessing could take a week or two, or more. What I meant is that the stickered pump should be something that you can get replaced through the RMA process. The lack of a response from AMD is undesirable, but it does not proove or suggest that a fixed, or at least improved, pump does not exist.

[Kailh]

That's not a different scale. Also, the GTAV graph ends at 60FPS, which may explain that difference.

yes it is. It visually makes the difference look smaller only ont he gta5 one. And why the fuck are benchmarks in a pump thread?

Also, look on the page for the benchmarks. It's some website I've never heard of. All of them there have it winning by a significant margin- if you look those games up elsewhere, NO where else has that.

[CommandMan7]

Really it was always a good card, people just expected too much from it and overhyped too much, alongside AMD overhyping it a bit.

This was exactly what I thought too, AMD got the hype train going a bit too fast and in the end it let people down.

[Techhog]

yes it is. It visually makes the difference look smaller only ont he gta5 one. And why the fuck are benchmarks in a pump thread?

Also, look on the page for the benchmarks. It's some website I've never heard of. All of them there have it winning by a significant margin- if you look those games up elsewhere, NO where else has that.

You're correct that the benchmarks don't belong here. That said, I can't say if there's anything wrong here or not due to the settings being used. TW3 benchmarks show the Fury X and 980 Ti tied with AA on and HairWorks off. I can't find a benchmark matching the settings in OP perfectly though. Moreover, the Fury X is actually closer to the 980 Ti in the GTAV benchmarks I've found, so if there's manipulation going on to make the Fury X look better, they're doing a lousy job of it.

[shawno]

yes it is. It visually makes the difference look smaller only ont he gta5 one. And why the fuck are benchmarks in a pump thread?

Also, look on the page for the benchmarks. It's some website I've never heard of. All of them there have it winning by a significant margin- if you look those games up elsewhere, NO where else has that.

Different test benches could cause this.

Why are you so offended by this? The fury wins some and loses some. Its still under performs the 980ti in the majority of titles

[Kailh]

You're correct that the benchmarks don't belong here. That said, I can't say if there's anything wrong here or not due to the settings being used. TW3 benchmarks show the Fury X and 980 Ti tied with AA on and HairWorks off. I can't find a benchmark matching the settings in OP perfectly though. Moreover, the Fury X is actually closer to the 980 Ti in the GTAV benchmarks I've found, so if there's manipulation going on to make the Fury X look better, they're doing a lousy job of it.

 

 

Different test benches could cause this.

Why are you so offended by this? The fury wins some and loses some. Its still under performs the 980ti in the majority of titles

well as far as the GTA5 one it runs better on nvidia for some reason, I don't expect it to be close at all. But if you look at every other website, they trade blows. Look at the site these came from, all fury x win aside from gta5 which they couldn't get away with claiming.

I'm so mad at fucking amd for this card. They had so many opportunities to do it right. The 295x2 had redundant asetek pumps. The fury x has....a cooler master one. The fuck amd. I'm also mad about the old hdmi that can't do 4k60 fps. Fuck htpc builds then, otherwise this card would have been perfect for the ultimate itx build, but it won't work with 4k TVs. Then there's the 4GB of memory. Their rebranded 290x has 8GB of memory, yet the new card has 4. Oh, and there's the changing the branding around to make the 390x be a 290x when it should have been like the 380x. Anyone not 100% in the loop with gaming hardware would think the 390x>290x and by a significant margin. It's stupid, AMD's choices hurt my head. Coming from the 295x2 where they nailed the first card to come stock with water cooling and have one of the best dual gpu cards in recent history to this is just cringe worthy.

 

Then there's the whole giving intel like no competition thing. Zen better be good.

[Techhog]

well as far as the GTA5 one it runs better on nvidia for some reason, I don't expect it to be close at all. But if you look at every other website, they trade blows. Look at the site these came from, all fury x win aside from gta5 which they couldn't get away with claiming.

I'm so mad at fucking amd for this card. They had so many opportunities to do it right. The 295x2 had redundant asetek pumps. The fury x has....a cooler master one. The fuck amd. I'm also mad about the old hdmi that can't do 4k60 fps. Fuck htpc builds then, otherwise this card would have been perfect for the ultimate itx build, but it won't work with 4k TVs. Then there's the 4GB of memory. Their rebranded 290x has 8GB of memory, yet the new card has 4. Oh, and there's the changing the branding around to make the 390x be a 290x when it should have been like the 380x. Anyone not 100% in the loop with gaming hardware would think the 390x>290x and by a significant margin. It's stupid, AMD's choices hurt my head. Coming from the 295x2 where they nailed the first card to come stock with water cooling and have one of the best dual gpu cards in recent history to this is just cringe worthy.

 

Then there's the whole giving intel like no competition thing. Zen better be good.

On the memory, they made the correct decision. The memory doesn't seem to be an issue in any of the 4k gaming benchmarks, so it seems that HBM and compression live up to the hype.

[Kailh]

On the memory, they made the correct decision. The memory doesn't seem to be an issue in any of the 4k gaming benchmarks, so it seems that HBM and compression live up to the hype.

there are many games that use more than 4GB, and much more if oyu are doing things like running a ton of mods. My issue with that stems form being future proof- I worry that if we do need more on average in the future, and you have  a fury x, and decide to add another or 2, the vram would be the only thing holding you back. Small thing, and nvidia is typically the one with less vram, and that's why it bothers me, that and the lower end card has more. BUT I'd like to think it had to do with the limitation of the first round of HBM chips.

[mr moose]

Different test benches could cause this.

Why are you so offended by this? The fury wins some and loses some. Its still under performs the 980ti in the majority of titles

 

Anyone who wants honesty in independent reviews should be offended that the scale changes depending on which card is performing better.   I don't generally agree with Kailth postings, but I do agree that those graphs:

 

1. are moot and add nothing to this thread

2. are presented in a biased way (changing the scale is presenting a bias).

3. don't seem to reflect the results of the other independent reviewers, thus indicating they are doing something different and not telling us what.

 

As consumers it is important that product representation is not misleading. especially from independent (or supposedly independent) reviewers.

[Techhog]

there are many games that use more than 4GB, and much more if oyu are doing things like running a ton of mods. My issue with that stems form being future proof- I worry that if we do need more on average in the future, and you have  a fury x, and decide to add another or 2, the vram would be the only thing holding you back. Small thing, and nvidia is typically the one with less vram, and that's why it bothers me, that and the lower end card has more. BUT I'd like to think it had to do with the limitation of the first round of HBM chips.

It is due to HBM1 limitations.

[Lethal Seraph]

Good good... looks like the competition is back. I'm looking forward for price drops for either and or nvidia.

Either way, no company doesn't owe you anything so being a fanboy for covering some company's unethical practices doesn't benefit us consumers.

[HashtagFTW]

Those benchmarks look pretty damn fake to me

[mr moose]

Those benchmarks look pretty damn fake to me

 

The numbers might be legit, but the huge problem is they are not presented fairly and they are not posted in any context (no explanation, link or or testing rational)

[HashtagFTW]

The numbers might be legit, but the huge problem is they are not presented fairly and they are not posted in any context (no explanation, link or or testing rational)

maybe, but pretty much all the reliable reviews have the Fury X beaten by the 980 Ti in W3 by anywhere from 3-5fps at 4k, and in this review, by a site that trys to make the Fury X look better, somehow the Fury X beats the 980 Ti by around 10%. A bit strange to be coincidental don't you think? Became pretty obvious to me when they didn't use Hairworks but used TressFX.

[shawno]

The way i see it is...

Sure the card is a bit of a let down in terms of performance but what it does is pave the way for future generations of cards. It gives both amd and nvidia a base on which they can improve.

As for the pump noise to get back on topic, it was not the fault of amd however the part they played would have been in the qa department. They would have heard it and should have addressed the issue prerelease.

[marldorthegreat]

It's not just the first batch. And way to chnge the scale on the benchamrks.

The one that you put it with the furyx loosing, a token of course, has a different scale to make it look better.

Typical AMD fanboy.

And cherry pick them.

Every site I've seen has w3 with the fury x loosing.

The pump is all them going CM instead of asetek

*Cough* Nvidia *cough* Fanboy *cough* 

[marldorthegreat]

The way i see it is...

Sure the card is a bit of a let down in terms of performance but what it does is pave the way for future generations of cards. It gives both amd and nvidia a base on which they can improve.

As for the pump noise to get back on topic, it was not the fault of amd however the part they played would have been in the qa department. They would have heard it and should have addressed the issue prerelease.

Its a let down now. But think about it. When the 290x launched it could not keep up with the 780ti, Now it is about the same if not better. AMD drivers will mean in about 3-6 months the fury x will be better than the titan x and 980ti.

[Sunako]

Its a let down now. But think about it. When the 290x launched it could not keep up with the 780ti, Now it is about the same if not better. AMD drivers will mean in about 3-6 months the fury x will be better than the titan x and 980ti.

what you are saying is pure speculation. There's no guaranty that it will become faster or that the 980 ti stays where it is. Its all what fis and will happens from you. We can't foresee the future so there's no point in claiming it here.

[Slinkey]

Same argument I hear all the time. The drivers will make it better. F--k that shit. When I buy a game I want it working here and now or at least a day zero patch.

Same thing when I buy a product, I want it working here and now.

[Khvarrioiren]

Same argument I hear all the time. The drivers will make it better. F--k that shit. When I buy a game I want it working here and now or at least a day zero patch.

Same thing when I buy a product, I want it working here and now.

 

So you never buy a new GPU? Pretty much every single GPU isn't working as well as it could out of the box after launch.