RX 480 is apparently killing pcie slots

bildad

Gawd
Joined
Jun 28, 2004
Messages
729
Who will do the confirmation? AMD? The users? The reviewers?
Confirmation is not going to come from AMD, so the debate on this topic will continue.

But the continued lost of AMDs market share will give the referendum against Polaris.

I'm sure nvidia has some high end power testing equipment. Maybe they can loan some out to testers. :D
 

Chimpee

[H]ard|Gawd
Joined
Jul 6, 2015
Messages
1,683
The card draws more power from the motherboard than is allowable by spec. And, it seems to impact the majority of cards. Every card tested at the PCI-E connector has shown this behavior, both review and retail samples. Show me ONE review that tested power at the PCI-E connector using a riser card that doesn't show PCI violation.

It's not a witch hunt. It's damage control by people who don't believe this.

Yup, and it would have been a non-issue if they use 8 pin power connector right from the get go, sure we will still bitch about power consumption but it would have been a non issue.
 

rezerekted

2[H]4U
Joined
Apr 6, 2015
Messages
3,043
The video on the HardOCP front page has nothing to do with the 480. WTF?

"DoNotPay Robot Lawyer Demonstration"
 

NKD

[H]F Junkie
Joined
Aug 26, 2007
Messages
8,869
This is just stupid. We are telling people to intentionally burn their cards lol.

Obviously this card is not controlling the power and not staying within tdp. That's the fuckin problem here. If you up the power limit and then run furmark, it's obvious it's gonna pull more. That is just stupid. Problem here is amd not stopping furmark from abusing the card.

Nvidia is smarter no matter what you toggle as soon as you run furmark it will throttle to stay under tdp.

AMD is just stupid for not doing the same. That's all they are stupid.
 

Track Drew

Limp Gawd
Joined
Dec 6, 2007
Messages
511
It would be interesting to see what the power numbers would look like if someone down-clocked a card by even as little as 5-10%. Increasing the clocks by small percentages dramatically increased power consumption for most reviewers. I'm guessing down clocking the card would improve power draw quite a bit.

You'll need Google translate:
AMD Radeon RX 480 im Test (Seite 12)

They undervolted the card, apparently it reduced system draw by 33 watts and increased performance (card could clock higher before reaching power target).

Obviously a poorly binned card would have less leeway here.
 

Daniel_Chang

[H]ard|Gawd
Joined
Jan 4, 2016
Messages
1,313
Yup, and it would have been a non-issue if they use 8 pin power connector right from the get go, sure we will still bitch about power consumption but it would have been a non issue.

Agreed. But oh well, the AIB cards shouldn't be effected. Now it's just a matter of waiting until they're out. My Intel HD4600 performs FAR worse than I expected. Can't wait to be able to properly game again.

This is just stupid. We are telling people to intentionally burn their cards lol.

Obviously this card is not controlling the power and not staying within tdp. That's the fuckin problem here. If you up the power limit and then run furmark, it's obvious it's gonna pull more. That is just stupid. Problem here is amd not stopping furmark from abusing the card.

Nvidia is smarter no matter what you toggle as soon as you run furmark it will throttle to stay under tdp.

AMD is just stupid for not doing the same. That's all they are stupid.

At STOCK the card goes over the limit. No adjusted power limit. No Furmark. At STOCK settings while GAMING the card is out of spec.
 

Ieldra

I Promise to RTFM
Joined
Mar 28, 2016
Messages
3,539
Yup, and it would have been a non-issue if they use 8 pin power connector right from the get go, sure we will still bitch about power consumption but it would have been a non issue.

Depends, if there is an issue with the power circuitry on the board, and the power draw is not distributed appropriately as a consequence, having an 8pin would not necessarily have helped the PCIe issue.
 

xorbe

Supreme [H]ardness
Joined
Sep 26, 2008
Messages
6,028
This is kinda bullshit. I feel like every time some new thing comes out it turns into a fight to belittle that item... This feels like this particular issue is on the next level and its Salem witch hunting levels. You guys are going ape shit over some possibly bad cards or mobos or even psus/eletricity, It will take some time before this is ever addressed as even a issue.

Yeah I don't know what to think ... it's like amd haters arrive en masse to fan the flames. Check my sig, I've rocked both Titans, not a single Radeon card in 10+ years. So it's only my observation. How many posters around here (1) have an RX 480 and (2) having issue with motherboard slot? Tom's shows about 80W average. Tom's noisy GPU power graphs 2-3 years ago is what caused me to buy an 860w power supply for my 430w PC, they've always been spiky, as far as I know. I'd trust Tom's over those clamp meters.
 

noko

Supreme [H]ardness
Joined
Apr 14, 2010
Messages
6,670
You'll need Google translate:
AMD Radeon RX 480 im Test (Seite 12)

They undervolted the card, apparently it reduced system draw by 33 watts and increased performance (card could clock higher before reaching power target).

Obviously a poorly binned card would have less leeway here.
Interesting:

Wonder if AMD can make a utility that tests a card out, stress testing, under volting and gives it a pass or fail? Redoes the power curves etc. keeps the clocks. Looks like AMD increase the voltage so more cards are available - if the case that is somewhat sloppy.
 

bildad

Gawd
Joined
Jun 28, 2004
Messages
729
Lot of placebo effect floating around. We should've been hearing about dead pci-e slots for years by now.

The difference is the 480 is an AMD designed and produced reference card, it's AMD's problem. The Asus 960 is Asus's custom design and problem, not Nvidia's or the 960 as a whole.
 

razor1

[H]F Junkie
Joined
Jul 14, 2005
Messages
10,120
Maybe the other way around lol,
j/k

Any case she just got a raise and possible bonus....
 

Araxie

Supreme [H]ardness
Joined
Feb 11, 2013
Messages
6,452
So, Who here is good with board design. Is something like this likely to be tweakable after the fact? Could they address where the power is pulled from using a firmware update to shift more power to the PCIe 12v connector, and less to the PCIe slot? Or is it more likely that they are going to have to recall these early boards?

God knows AMD can't afford to be spending their limited cash to fix a mess like this right now. I'm certainly hoping this can be affordably fixed.

Safer option would be recall all the cards and they manually flash the BIOS if and only IF the card allow through BIOS total control of power balance.. Nvidia cards allow that, easily. AMD cards I used doesn't, in my case my HD7970, 280x, 390 and 390X from different vendors none allowed to power balance and power tables edition, so it's possible they have dedicated advanced triac transistor, micro-relays and micro-mosfet-switch to balance the load "mechanically", in that case the option would be a complete redesign because even with a 8Pin power connector the card will split 50/50 the load between PCI-E port slot and the PCI-E power connector, so screwed anyway, so I just hope they can manage the power control balance via software through BIOS so they can, probably in the best case scenario they could make an in-house tool to flash BIOS.. that's really hoping the best case scenario, worse case then they will have to limit via driver and hidden for the user a way to limit the clocks of the cards for about 20W but that's just A LOT of MHZ less as razor1 said above about 10% less clocks would be fine to keep the card at better load margins, this would entirely force AMD to change the base and boost clocks from every piece of marketing and the official specification of the reference card so they will be again bashed as liars. in my opinion they are screwed.

And for everyone thinking this can be fixed via driver, stop with the delusion there please... this change have to be directly on the card. Physically or via BIOS at best..
 

Ieldra

I Promise to RTFM
Joined
Mar 28, 2016
Messages
3,539
The difference is the 480 is an AMD designed and produced reference card, it's AMD's problem. The Asus 960 is Asus's custom design and problem, not Nvidia's or the 960 as a whole.
To be fair, nvidia have to approve the board design, so it's on them as well
 

primetime

Supreme [H]ardness
Joined
Aug 17, 2005
Messages
7,167
Im absolutely sure it was some kind of Dumb Ass marketing decision to use the 6 pin connector. I truly hope whoever made this call is FIRED this week at AMD no matter who it was - It was Clearly a Dumb fucking decision any kind of way we look at it. Someone mentioned they could have used the 8pin just for the headroom alone....or they could have added a second optional connector for over clockers (even though it needs 2 for stock clocks)
 

Schmave

[H]ard|Gawd
Joined
Jan 2, 2001
Messages
1,720
You'll need Google translate:
AMD Radeon RX 480 im Test (Seite 12)

They undervolted the card, apparently it reduced system draw by 33 watts and increased performance (card could clock higher before reaching power target).

Obviously a poorly binned card would have less leeway here.
That's really interesting. It definitely looks like the chips were designed to run at lower clock speeds and/or voltages. On the PCPer interview with Raja and that other guy the other day they did make some off-hand comment about the WattMan utility that you can not only use it for overclocking, but to reduce power consumption. Maybe they had seen this issue before.
 

rbuass

n00b
Joined
Jul 1, 2016
Messages
29
Need to run a retest at stock power limit to be completely valid. (Nothing personal, just a peer observation that I would personally do)
We have tested for more than 5 hours alive.
But for 5 or 7 minutes vídeo, also with explanation, we cannot show a summary this way.
Is quite obvious, that we need to show usual 50% PL, that's normal for the public.
If we have dozens of situation....

Idle...
No PL light game
No PL heavy game
pl 25 and pl 50 in heavy games
no PL, pl 25 and pl 50 furmark
Overclocking 3, 4 and 5% (max we have reached)
So, more explanation, what we are talking about...

So, if we consider dozen situations + explanation, and we will show honest preparation (no cheat), ... for 2 minutes to show settings and blablabla... will have about one hour video... and will be quite boring.
Then is the reason that we used just 2 situation, and sorry if is not enough to prove that the card request is above specifications.
That's for who believe will help... if for some reason people don't believe in our point of view, we just talk that's not our intention.
 

MN Scout

Supreme [H]ardness
Joined
Jan 26, 2003
Messages
4,713
128 from the power supply PLUS 126 from the PCI Express connector when running the Furmark benchmark. 254 watts total. I thought this was supposed to be a card that didn't use much power.
 

Ieldra

I Promise to RTFM
Joined
Mar 28, 2016
Messages
3,539
We have tested for more than 5 hours alive.
But for 5 7 minutes vídeo, also with explanation, we cannot show a summary this way.
Is quite obvious, that we need to show usual 50% PL, that's normal for the public.
If we have dozens of situation....

Idle...
No PL light game
No PL heavy game
pl 25 and pl 50 in heavy games
no PL, pl 25 and pl 50 furmark
Overclocking 3, 4 and 5% (max we have reached)
So, more explanation, what we are talking about...

So, if we consider dozen situations + explanation, and we will show honest preparation (no cheat), ... for 2 minutes to show settings and blablabla... will have about one hour video... and will be quite boring.
Then is the reason that we used just 2 situation, and sorry if is not enough to prove that the card request is above specifications.
That's for who believe will help... if for some reason people don't believe in our point of view, we just talk that's not our intention.
If you do a more in depth article please link it here
 

Araxie

Supreme [H]ardness
Joined
Feb 11, 2013
Messages
6,452
We have tested for more than 5 hours alive.
But for 5 or 7 minutes vídeo, also with explanation, we cannot show a summary this way.
Is quite obvious, that we need to show usual 50% PL, that's normal for the public.
If we have dozens of situation....

Idle...
No PL light game
No PL heavy game
pl 25 and pl 50 in heavy games
no PL, pl 25 and pl 50 furmark
Overclocking 3, 4 and 5% (max we have reached)
So, more explanation, what we are talking about...

So, if we consider dozen situations + explanation, and we will show honest preparation (no cheat), ... for 2 minutes to show settings and blablabla... will have about one hour video... and will be quite boring.
Then is the reason that we used just 2 situation, and sorry if is not enough to prove that the card request is above specifications.
That's for who believe will help... if for some reason people don't believe in our point of view, we just talk that's not our intention.

Hello and thanks for the answer, but precisely talking about every different possible scenario the first one that had to be taken in consideration was the utilization first the card at stock settings, out of the box and nothing else. and then you could play and toy with whatever setting requested by the users if wanted. but it's the stock settings what truly have the attention of the people, that would be the best approach in my opinion, start from stock settings.
 

xorbe

Supreme [H]ardness
Joined
Sep 26, 2008
Messages
6,028
128 from the power supply PLUS 126 from the PCI Express connector when running the Furmark benchmark. 254 watts total. I thought this was supposed to be a card that didn't use much power.

No way is that little card sustaining 254w like an og Titan, gpu would pop with that cooling solution.
 

Sw1sher

Limp Gawd
Joined
Mar 15, 2016
Messages
142
Everywhere i go, i keep seeing comments about how the 480 is a disappointment, didnt meet the hype, etc. But damn, the 480 blowing shit up now? AMD really causing a ruckus
 

yourgrandma

[H]ard|Gawd
Joined
Apr 15, 2011
Messages
1,386
It's ironic this budget card has trouble with low end motherboards, not that i would want to torture test a costly high end motherboard with one, but knowing this info makes me not want to recommend it anyone. But I'm rooting for them to get it fixed quickly at least before nvidia unveils the 1060.
Personally, I think this is all a part of AMDs master plan to blow up the PCs of cryptocurrency miners :p
I wish, is the rx 480 any good at mining? I'd figure we'd hear problems from the miners first if this is a major issue or at least when the 4gb model is actually obtainable.
 

Daniel_Chang

[H]ard|Gawd
Joined
Jan 4, 2016
Messages
1,313
Everywhere i go, i keep seeing comments about how the 480 is a disappointment, didnt meet the hype, etc. But damn, the 480 blowing shit up now? AMD really causing a ruckus

I don't consider it a disappointment. It offers 390 performance at 380 price and power draw. It's outstanding. The only problem with the card right now is the PCI-E power draw. Easy solution - don't buy the reference card.
 
D

Deleted member 93354

Guest
Here's what will happen:

AMD will lower p-state claim it as a fix. Performance will suffer as boost clocks will throttle more often. Overclock will be disabled. (Cost AMD nothing)

People will bitch up a storm, AMD will shrug their shoulders in a "mia culpa" and ignore the issue similar to how NVIDIA boards were melting the MBGA solder points and malfunctioning (Basically "It cost it too much, so F you")

Some lawyers will get involved eventually smelling blood. They will get rich and everyone will get a $10 coupon that bought one of these past a certain date (The confirmation there is a problem in an official press release or letters sent to owners)
 

Araxie

Supreme [H]ardness
Joined
Feb 11, 2013
Messages
6,452
No way is that little card sustaining 254w like an og Titan, gpu would pop with that cooling solution.

oh yes it can, the same way lot of 980TI were able to sustain above 450W.. I have some cards with custom BIOS using ~480W ALONE at about 1625mhz. =) so yes can be tolerated, it's harm? is good? of course isn't good but nobody knows for how much time, you don't stay 24/7 running furmark with a RX 480 to reach those numbers. the guy who is aimed to pick this card, will turn on his machine, check FB, check Twitter, watch some Pron, fire up couple of games for couple of hours and done.

Everywhere i go, i keep seeing comments about how the 480 is a disappointment, didnt meet the hype, etc. But damn, the 480 blowing shit up now? AMD really causing a ruckus

just see it as free marketing AMDBelievers will say "hey everything said in forums in videos and techsites it's fake im gona buy anyway those cards" Nvidia guys will be still attacking the card and promoting the waiting time for the GTX 1060, in any sense it's free marketing for AMD as it generate more interest on the card (not matter as being good or bad) marketing is as good as bad as it can be, it's still marketing "our card it's fantastic it have reached historic amounts of search on the web, people are really interested in our new GPU" see? that's how work..
 

Ieldra

I Promise to RTFM
Joined
Mar 28, 2016
Messages
3,539
oh yes it can, the same way lot of 980TI were able to sustain above 450W.. I have some cards with custom BIOS using ~480W ALONE at about 1625mhz. =) so yes can be tolerated, it's harm? is good? of course isn't good but nobody knows for how much time, you don't stay 24/7 running furmark with a RX 480 to reach those numbers. the guy who is aimed to pick this card, will turn on his machine, check FB, check Twitter, watch some Pron, fire up couple of games for couple of hours and done.



just see it as free marketing AMDBelievers will say "hey everything said in forums in videos and techsites it's fake im gona buy anyway those cards" Nvidia guys will be still attacking the card and promoting the waiting time for the GTX 1060, in any sense it's free marketing for AMD as it generate more interest on the card (not matter as being good or bad) marketing is as good as bad as it can be, it's still marketing "our card it's fantastic it have reached historic amounts of search on the web, people are really interested in our new GPU" see? that's how work..
Have you Googled 'rx480' recently?
 

NKD

[H]F Junkie
Joined
Aug 26, 2007
Messages
8,869
Come on leldra...We know thats your other brand new account! lol;) jk...but yea people with 4 post making big claims are almost always trolls

yep and its funny he made the thread responded once and disappeared for hours. If this card killed my board I would be cursing out amd in every thread and everywhere until I got a result. I think someone asked him to post pictures too lol.
 

primetime

Supreme [H]ardness
Joined
Aug 17, 2005
Messages
7,167
POWERCOLOR-Radeon-RX-480-DEVIL.jpg


Is this dam thing using a single 6 pin? lol
 

Dan

Supreme [H]ardness
Joined
May 23, 2012
Messages
7,906
It's ironic this budget card has trouble with low end motherboards, not that i would want to torture test a costly high end motherboard with one, but knowing this info makes me not want to recommend it anyone. But I'm rooting for them to get it fixed quickly at least before nvidia unveils the 1060.

I wish, is the rx 480 any good at mining? I'd figure we'd hear problems from the miners first if this is a major issue or at least when the 4gb model is actually obtainable.

Been running 2 in crossfire doing fur mark while I'm working on my higher end Amd system. So far so good.
 

NKD

[H]F Junkie
Joined
Aug 26, 2007
Messages
8,869
What the fuck? I have doubts that that's at stock settings.

I think its at 50% power thing. enabled. Idk whats the point. Only purpose here is to see how bad it is in controlling tdp. I mean you run furmark and allow more power its gonna do just that. I think amd is to blame here for being stupid and letting this happen. Why would you not control the TDP under furmar, lock that shit down. Nvidia does that just fine.
 
Top