Discussion: Polaris, AMD's 4th Gen GCN Architecture

[gamerk316]

I will say this: There's enough reports out there where AMD should put out an official response. This is obviously NOT the type of launch hype there were hoping for, and this could get really bad for AMD really quickly if not addressed.

 

[Ags1]

I signed up to the AMD technical forums. There were only three RX480 cases related to the power draw issue - two were requests for clarification, of was an OC'd 3-way XFire setup.

EDIT: No, there are few more threads asking for clarification. No AMD response on any of them.

 

[BurgerandChips66]

Here is a sneak peak of the new RX 480 reference card (so V2).

There are additional power connectors and an improved cooler to allow increased overclocking and hopefully avoid issues with excessive PCIe slot power consumption.

Hope it still retails for $199.....

😉

 

[Robert Cook]

Here is a sneak peak of the new RX 480 reference card (so V2).

There are additional power connectors and an improved cooler to allow increased overclocking and hopefully avoid issues with excessive PCIe slot power consumption.

Hope it still retails for $199.....

😉

That is a pro duo, not a RX480.
EDIT: wait sarcasm?

 

[Bem-xxx]

I signed up to the AMD technical forums. There were only three RX480 cases related to the power draw issue - two were requests for clarification, of was an OC'd 3-way XFire setup.

Only three cases vs thousands of RX480 already sold.

 

[Mousemonkey]

AdoredTV did a video that shows this issue is not limited to the RX480 but multiple other cards including some by the green team. Is this problem being blown out of proportion?

https://youtu.be/kFuYc2FHgjw

Did you read that GTX 960 review? I don't think you did because if you did you would see that was a review of multiple AiB cards and that only the Asus GTX 960 Strix had that spiking issue, none of the other ones did.

I have two GTX960's and both of them run just fine regardless of what motherboard I run them in, and I have four to choose from.

 

[Ags1]

I signed up to the AMD technical forums. There were only three RX480 cases related to the power draw issue - two were requests for clarification, of was an OC'd 3-way XFire setup.

Only three cases vs thousands of RX480 already sold.

I thought that was obvious. Yes, only one actual report of a smoking motherboard as far as I can see on the AMD support forum. Plenty of cases asking for reassuruance - with no response from AMD.

 

[BurgerandChips66]

Yes, it was a joke..... I would have edited out the second GPU backplate but it's late.... !

 

[turkey3_scratch]

Bem-xxx, it is clear to me you are an owner of one of these cards. Are you not? I can tell by the way you are getting very defensive over this and trying ti tailor evidence and sources and claim other sources to be illegitimate. Are you going to deny Tomshardware's testing on the RX 480? Are you going to deny the people having problems? This card has only been out two days now, and six or more issues online seems like a lot to me, not to mention those people who don't participate in Internet forums or haven't spoken of their own issues.

 

[Bem-xxx]

I signed up to the AMD technical forums. There were only three RX480 cases related to the power draw issue - two were requests for clarification, of was an OC'd 3-way XFire setup.

Only three cases vs thousands of RX480 already sold.

I thought that was obvious. Yes, only one actual report of a smoking motherboard as far as I can see on the AMD support forum. Plenty of cases asking for reassuruance - with no response from AMD.

AMD should pursue legally the Nvidia's astroturfers.
https://consumerist.com/2006/02/06/did-nvidia-hire-online-actors-to-promote-their-products/

 

[turkey3_scratch]

I signed up to the AMD technical forums. There were only three RX480 cases related to the power draw issue - two were requests for clarification, of was an OC'd 3-way XFire setup.

Only three cases vs thousands of RX480 already sold.

I thought that was obvious. Yes, only one actual report of a smoking motherboard as far as I can see on the AMD support forum. Plenty of cases asking for reassuruance - with no response from AMD.

AMD should pursue legally the Nvidia's astroturfers.
https://consumerist.com/2006/02/06/did-nvidia-hire-online-actors-to-promote-their-products/

Why are you pointing out 2006 articles of questioning legitimacy?

 

[Bem-xxx]

Bem-xxx, it is clear to me you are an owner of one of these cards. Are you not? I can tell by the way you are getting very defensive over this and trying ti tailor evidence and sources and claim other sources to be illegitimate. Are you going to deny Tomshardware's testing on the RX 480? Are you going to deny the people having problems? This card has only been out two days now, and six or more issues online seems like a lot to me, not to mention those people who don't participate in Internet forums or haven't spoken of their own issues.

I have a R9 390 CrossFire, and I dislike Nvidia's shady marketing practices, because history repeats itself.

 

[turkey3_scratch]

Bem-xxx, it is clear to me you are an owner of one of these cards. Are you not? I can tell by the way you are getting very defensive over this and trying ti tailor evidence and sources and claim other sources to be illegitimate. Are you going to deny Tomshardware's testing on the RX 480? Are you going to deny the people having problems? This card has only been out two days now, and six or more issues online seems like a lot to me, not to mention those people who don't participate in Internet forums or haven't spoken of their own issues.

I have a R9 390 CrossFire, and I dislike Nvidia's shady marketing practices, because history repeats itself.

Well, I respect your opinion of course, but what about AMD's marketing? What are these shady Nvidia marketing practices? Also, why does it matter to you how a product is marketed, as a consumer I just want to get my hands on whatever is best.

 

[-Fran-]

Bem-xxx, you need to Keep Calm and Wait for More Information.

Read the same things you posted and try to analyze each one and see why they don't provide a counter argument to AMD's problem with the RX480.

Cheers!

 

[jimmysmitty]

The RX 480 is selling a lot. Multiple reports done by who? One is a troll who has been banned. If the issue was real, would be a hundred of reports, not two or three made by the same people with different nicknames.

Selling a lot? There has been low inventory to start.

Second what proof do you have that the reports are from the same person or they are a troll?

Bem-xxx, it is clear to me you are an owner of one of these cards. Are you not? I can tell by the way you are getting very defensive over this and trying ti tailor evidence and sources and claim other sources to be illegitimate. Are you going to deny Tomshardware's testing on the RX 480? Are you going to deny the people having problems? This card has only been out two days now, and six or more issues online seems like a lot to me, not to mention those people who don't participate in Internet forums or haven't spoken of their own issues.

I have a R9 390 CrossFire, and I dislike Nvidia's shady marketing practices, because history repeats itself.

Personal opinion vs facts. I can easily point out any company that has had "shady" marketing, AMD, Intel or any company, easily because they all at some point in time market poorly. Or do you not remember the Bulldozer marketing that got the entire marketing team fired?

The facts are that Toms found an issue as did other sites. You are letting your personal feelings interject and using cherry picked examples to bash nVidia when the fact is that this is not AMD vs nVidia, this is a matter that there is an issue with the RX 480 pulling more power from the PCIe slot than is specced and is causing issues. AMD has even admitted they are working on a possible software issue:

http://forums.anandtech.com/showthread.php?p=38330571

So is nVidia paying AMD to say they know about the issue and are looking to fix it?

And look at what I posted earlier. Even the vastly more power hungry Fury X doesn't hit over 33W spikes for power draw from the slot.

 

[Mousemonkey]

I signed up to the AMD technical forums. There were only three RX480 cases related to the power draw issue - two were requests for clarification, of was an OC'd 3-way XFire setup.

Only three cases vs thousands of RX480 already sold.

I thought that was obvious. Yes, only one actual report of a smoking motherboard as far as I can see on the AMD support forum. Plenty of cases asking for reassuruance - with no response from AMD.

AMD should pursue legally the Nvidia's astroturfers.
https://consumerist.com/2006/02/06/did-nvidia-hire-online-actors-to-promote-their-products/

Why are you pointing out 2006 articles of questioning legitimacy?

http://alienbabeltech.com/old_abt/viewtopic.php?f=44&t=24247&start=2160

 

[turkey3_scratch]

I don't know what alienbabeltech is... I try not to go to websites I'm unfamiliar with, don't feel like booting up Linux right now. Sorry.

So does anybody think this issue could be fixed with drivers?

 

[BurgerandChips66]

https://www.youtube.com/watch?v=nZYHx_tad3U

GTX 970 SSC vs. RX 480 ..... of interest were the DX12 benchmarks. AOTS was not tested by the way.

 

[turkey3_scratch]

https://www.youtube.com/watch?v=nZYHx_tad3U

GTX 970 SSC vs. RX 480 ..... of interest were the DX12 benchmarks. AOTS was not tested by the way.

It was tested in the Tomshardware review.

 

[-Fran-]

http://alienbabeltech.com/old_abt/viewtopic.php?f=44&t=24247&start=2160

I don't... Get all that rambling in that thread.

Mind explaining? I clicked the 1st link inside the link you provided and the guy posting actually predicted HBM correctly. Although the other information he gives is dubious. I do know AMD was behind GDDR3 and GDDR5, but it was as part of the JEDEC consortium. That is why the 4870 was the first one with it.

Cheers!

 

[Mousemonkey]

http://alienbabeltech.com/old_abt/viewtopic.php?f=44&t=24247&start=2160

I don't... Get all that rambling in that thread.

Mind explaining? I clicked the 1st link inside the link you provided and the guy posting actually predicted HBM correctly. Although the other information he gives is dubious. I do know AMD was behind GDDR3 and GDDR5, but it was as part of the JEDEC consortium. That is why the 4870 was the first one with it.

Cheers!

You've got go digging, ABT were one of the sites that got approached by AMD to promote their products in return for samples and they refused. It kind of went downhill from there.

 

[-Fran-]

You've got go digging, ABT were one of the sites that got approached by AMD to promote their products in return for samples and they refused. It kind of went downhill from there.

Ah... That explains it then. I think I remember something about that...

Cheers!

 

[Rogue Leader]

It's safe to say it is indeed a problem that should be addressed by both AMD and board partners before their own cards come out. Even Rogue Leader says he can notice through the speakers. Though with his optimism nothing seems to be able to disappoint him!

I lived through computer gaming in the late 80's and 90's, it takes a lot to disappoint me. You used to buy a GPU and spend a day or 5 beating the crap out of Windows and DOS getting your games to not crash with it. ATi, 3DFx, and the rest were like "yeah we'll fix that, some day...." This IMO is a joke in comparison.

Not only that performance wise I'm happy with it. I think a lot of people here could agree if not for the power problem this is a pretty great value at $200. It does Firestrike in the same range as a pair of cards I spent about $300 on, and even with the power issues uses half the power. Ignoring my setup its $100 cheaper than the nearest thing that can perform to it. Everything I've played works excellently. I wasn't gonna bother buying Doom, then I tried the demo and it works perfect on ultra, I'm gonna buy it now (and its on Steam sale!).

Another thing I want to point out is I have not overclocked it or touched anything really (and I don't plan to, with my CPU there is not much to gain). I have run firestrike on it probably 15 times and played a few hours worth of games. No I didn't play Witcher for 7 hours straight, but I do have a 5 year old mid range motherboard and there has been no indication of any damage. That said its something that will happen without warning.

I do believe there is a power problem, however I have a hard time believing everyone on the internet who says they damaged their motherboard, especially those with higher end 990FX boards. Unless they were overclocking which we have all clearly figured out is a bad idea right now. Of course nobody would admit to that, because if theres a chance that AMD will cover people for motherboards, I'd bet overclocking would preclude you from that deal. I had a problem with the voltage locks and overclocking on one of my old cards, I contacted XFX and while they print overclocking all over the box, the tech very clearly told me he can't help me and overclocking voids the warranty.

So I find it curious that I have not eaten a motherboard yet, and TBH I am gonna keep playing this thing and see what happens. I would say its likely if a board is gonna fail, a 5 year old mid range board would have a better chance of failing. Its very possible at stock clocks this isn't that big of a deal, it still needs to be fixed, that background noise through the speakers concerns me and if it kills my now out of production X-Fi Titanium HD I will be bent, but the motherboard I don't care about. I don't know but we will see....

 

[turkey3_scratch]

Well, Tomshardware testing showed it to be around 82W average for the PCIe slot, only 7W more than the rating. I would have to imagine that the temperature inside your case also play a big role. It is simple heat addition, you know? If case temps are cool on the inside, the issue is probably less realistic or not realistic at all. You're probably also right about those people overclocking their RX480s, those people who have had problems.

Could you explain to me, though, why it in particular affects noise through the speakers? What does PCIe slot power for a GPU have to do with noise output through dpeakers?

 

[Rogue Leader]

Well, Tomshardware testing showed it to be around 82W average for the PCIe slot, only 7W more than the rating. I would have to imagine that the temperature inside your case also play a big role. It is simple heat addition, you know? If case temps are cool on the inside, the issue is probably less realistic or not realistic at all. You're probably also right about those people overclocking their RX480s, those people who have had problems.

Could you explain to me, though, why it in particular affects noise through the speakers? What does PCIe slot power for a GPU have to do with noise output through dpeakers?

I do have very good case cooling, I have 2 200mm intake fans (front and side), and 1 200mm exhaust at the top and a 120mm at the back. Even with that sound card blocking the fans on my old GPU I never had problems, thanks to all my fans, it helps a lot.

So now as to my speakers, I have a PCIe sound card in the slot next to the GPU, so I would assume that excessive power draw is causing some interference in the card. Its funny because Creative put a whole shield around the card to prevent against this, but clearly it didn't work. Also I'm not sure with cheap speakers you would detect this, I run a pair of M-Audio studio monitors which are very sensitive and have more range than a normal pair of computer speakers. When you spend $200 on a sound card, you tend to want to make sure you get your moneys worth out of the sound.

PS the interference is VERY minor, I need to put my head next to the speaker to clearly hear it, I only looked for it because I heard of folks having onboard sound issues.