AMD Gives Statement on the PCI-Express Overcurrent Problems

It's very possible that all this mess was caused by some AMD PR exec decisions to make 480 appear as the power efficient champion using only a 6 pin power connector and at the same time increase the GPU clocks beyond what is reasonable with this power supply.

Now the driver team (software) with the help of engineers (hardware) must clean this mess lowering the sold GPUs power needs...and performance via drivers.

A better option could be to flash a modified bios but this is not a normal (and secure) operation for end users.

In a perfect world all AMD 480 GPUs will be recalled and substituted with at revised GPU with at least a 8 pin power connector.We don't live in a perfect world and GPUs are not cars. :nerd:

It will be interesting to compare the performance of custom PCB 480 models with 8 pin connectors and better cooling (Sapphire RX 480 Nitro...) VS the reference model at launch VS the reference model with limited power needs in a next drivers version.

We need some AMD GPUs success (even Nvidia users): real competition is needed to lower prices and get faster hardware improvements.

 

How much power does an 8 Pin provide?

 

150W. Custom 480's with 2x6pin or 1x8pin and problem is gone. Anyone knows in that thread (Asus and MSI teases with custom 480) what kinda power delivery it has?

 

Wow what a clusterf$%k. Did they not test these cards properly on various motherboards before releasing them.

If they do downclock the 480 to fix the issue, then i agree with others here in that all benchmark results so far will be invalid, and they will need to retest them again after the driver fix for more accurate results.
Also i would imagine that overclocking or running the 480's in Xfire would only worsen this power draw issue, so maybe best to run a single 480 and not overclock them.

Also where are all the AMD guys now, the ones claiming in the past few days that there was no issues with the excess power draw? Vase??
AMD just pushed to hard with this gpu, there was to much pre-release hype that caused a lot of disappointment once it was released.
And now these power draw issues and the possibility of having the 480 downclocked to fix this, the 480 is not a good release at all!

 

Not only does it take more power than a 1070, it's asking for more than 75W over PCI lane?!


I really want this card to get AMD back in the game, come on !

 

yeah dont worry mr. obtrusive.
i am still here and no issues with dead mainboards have arisen.
why would benchmark results be invalid?
the card works perfectly with the highest performance on all standard modern mainboards.

i dont see ANY mainboard section of a forum flooding with posts of dead boards or ANYTHING like it.
i just see worthless stuff being posted by users on here that like to produce chaos.

there have yet to be reports about any ISSUE apart from one guys youtube who deliberately tested the card on a motherboard from 2010 that doesnt even have windows 10 or windows 8.1 drivers available

VIDEO:
https://www.youtube.com/watch?v=rhjC_8ai7QA

MOBO:
-> http://www.foxconnchannel.com/ProductDetail.aspx?T=motherboard&U=en-us0000493

Spoiler

half of the people here were playing with their nintendo game boys or sega game gears when this came out.

and what did the board do? power down the card, as spec dictates.

all is fine. it is a no issue.
amd will have to deal with the mess the internet-troll ****storm created.
by probably undervolting and leaving boost clocks or by reducing power level via an optional driver profile hotfix ... and in the same time they will comment:

"these adjustments have been made to assure 100% that it will run on ALL motherboards with a single pcie 1.0 slot from 2005-2016 that wasnt even designed for gamer pcs.
-> people who have standard mainboards that are newer than from 2010 or high end mainboards can just set the powerlevels back to xy as they were before or dont have to install the hotfix at all unless they experience shutdowns of the graphics card.
we had to do this because public opinion was misguided by toxic individuals
thank you for your understanding."

or maybe they will come up with some even more elegant solution, whatever.
with it being a non-issue still, the way of the "solution" is absolutely irrelevant.

but then at least there will be quiet faces and the trolls can actually get back to doing homework or something like that.

 

To be fair, from what I see often is that AMD just delivers the usual hype almost no different than Nvidia, but it is the AMD cheerleading squads on the forums that actually hype up the products beyond any possibly reasonable level. So when the actual product is delivered with a realistic goal and performance, everyone is let down from all the hype. So you have the AMD cheerleading squad trying to pretend the product meets the hype, while you have everyone else call bullsht on the hype.

It's not going to be any different with Zen or Vega.

 

Little hype is always welcome but AMD pruducts get so much hype it hurts them much more than helps.

Take a look at gtx1060, no one cares, one barely talks about it.

 

The benchmarks should indeed be invalidated (obviously not the reviewers fault) if they are to be changed and/or are inaccurate.

 

Because it's Nvidia, so no one is pretending that they need a second coming of christ. The card will perform how it performs, and will get slotted into its area of the market and sell well. No need for excuses or hype.

 

'If' they did need to downclock or change power settings, then it will change the bench results, temperature results etc, that's why!

Also if the 480's are working perfectly as you claim, then why would AMD release a statement about it and need to release a driver fix to correct the power draw issues?

Even if people are running older motherboards, they still have the right to buy a new gpu that works, and one that is not out of spec in regards to excess power draw.

 

a non-issue (or respectively a non-issue that is depicted as an issue by a reviewer minority 3 out of over 200, at the same time by a community majority -> nvidia users -> NEGATIVE hype = hatespeech = slander/libel/defamation) doesn't disparage the 480x performance achievements. "sorry to burst your bubble" (as the toxic user of these forums would insert here)
so all the hype was completely correct and right. we are going to get overclocks up to 1600mhz with custom cooling, confirmed by non-amd friendly kyle from [H]
and with crossfire 480 we get better than single 1070 performance in many games that scale (confirmed by techpowerup review and by hardwareunboxed)
so NO, the hype wasnt wrong as you try to put it in your general negative attitude. the hype was correct. and all thats being done now is people being copycats and trying to make mountains out of molehills.
LITERALLY MOLEHILLS (power spikes in pcie-slot drain).
that existed on several other cards.

 

This guys hype certainly wasn't correct!

 

i'll give you a rereview.

undervolted @154W max:

Spoiler

vs not undervolted @ 166max from guru3d review

Spoiler

now re-review your logical sense.

a re-review makes no sense if just driver parameters that are settable by user change.

you all ask for a re-review while it isnt even CLEAR if there will be this type of solution. but all say "ohhh downgrade incoming".

even if it WOULD be an underclock or power level thing.
how is it a downgrade when 99,99% of the users will just use the same power level as before, because they didnt have a problem in first place.

 

A perfect example of the cheerleading squad that is doing more harm than good for AMD on forums....

How about you chill? Seriously, sit back and relax, it's a fcking piece of silicon we are discussing here, not your children. The RX480 has a problem right now, as do nearly all products as complicated as it, that will be fixed. Period, no need to be in denial about it and go on ranting....

And no, Crossfire or SLI will never be considered a reasonable alternative to a single quality card just simply due to the barriers of entry. Namely, power usage is now doubled compared to the single card,so is the heat increased in the rest of the system, noise goes up, you need an SLI/Crossfire ready motherboard which costs more; then after all this, you need to deal with the nightmare driver situation for SLI that turns into hell by the time Crossfire is approached (as evidenced in the majority of games tested for the RX480 Crossfire not receiving any performance boosts).

So like I said earlier, chill. The RX480 is a nice card that at $200 is very attractive. No one is saying otherwise. Stop trying to make the card be a cure for cancer...

 

so what? did he hurt anyone with that, or did he hurt you? so you need to be all toxic now?
he saw the positive side of things. i dont know if he was right or wrong. it doesnt matter. a hype is called hype for what it is.
but hatespeech is also called hatespeech for what it is.
i will never hatespeech nvidia products because users overhype them.
why would i? only really narrow-minded people would hatespeech products that are overhyped by others.

 

vase mate, you're making yourself look a bit foolish. Seriously, just stop.

 

i will reduce my posts down to almost single sentences now because it just seems longer passages overwork some people.

i can tell you what's in it for me to make you accept the facts.
-> helping out just makes me feel better.

i just wanna know why you lie. why do you lie? what's in it for you personally to lie about crossfire relevance & performance?

 

why is it foolish to show that RX480 crossfire is 14% faster in listed triple A games than a single
1070?
explain me how you come to this conclusion. just explain:

"ok vase it is foolish to show that crossfire is 14% faster because..."

contine the sentence please

 

I'm not going to drag myself down to your level. I'm smarter than that.