RX 480 fails PCI-E specification

overclock.net/t/1604477/reddit-rx-480-fails-pci-e-specification/100

has a collective board ever been as BTFO as Sup Forums right now?

Other urls found in this thread:

reddit.com/r/Amd/comments/4qmlep/rx_480_powergate_problem_has_a_solution/
tomshardware.com/reviews/geforce-gtx-750-ti-review,3750-20.html
teledynelecroy.com/doc/docview.aspx?id=7649
pcper.com/reviews/Graphics-Cards/Power-Consumption-Concerns-Radeon-RX-480
en.wikipedia.org/wiki/PCI_Express#Power
twitter.com/NSFWRedditGif

>reddit

>he trusted poo in loo man

LITERAL HOUSEFIRE
Can't wait for the first deep fried motherboard

The 960 had spikes up to 250W, higher than the 480. Where was the outrage and fried motherboards? Why is it okay when Nvidia do it, as per usual?

>ushanka
why do AMD get the best hat?

>b-b-b-but nvidiaaaaa

Nvidia has more money to silence the press
im sure they gave an extra tip to reviewers about the RX480's power consumption

that would be so scummy, we shouldn't find out until someone burns their house down instead

That was a problem with ASUS-specific implementation of the 960, this is the reference card from AMD.

Spikes up to, but doesn't average at. Problem with the 480 is its averaging well above PCI specification in stock configuration. Overclock on a midrange board and enjoy the fireworks.

So is that shitty ASUS 960. But you'd have to be retarded to buy ASUS so that's okay.

>has a collective board ever been as BTFO as Sup Forums right now?
>I just come here to shill
FTFY, fucking shill

spikes? its spending more than half its time above 150W when the AMD card is not even hitting that. thees graphs are misleading.

here we go again

reddit.com/r/Amd/comments/4qmlep/rx_480_powergate_problem_has_a_solution/

DELET THIS

Already proven false. Move along everybody

A disaster !!

I don't know why everyone is shitting the bed so hard about this, as long as you aren't using some cheap piece of shit foxconn or equivalent board you'll be fine. I turned up the pci-e power draw in my custom bios for my 970 up to 150w, no issues at all in the past year. This whole issue is just shills shilling shills.

glad I bought a 1070

DELET THIS

Good counterpoint to shills in there:

tomshardware.com/reviews/geforce-gtx-750-ti-review,3750-20.html

When nvidia 750ti drew 125w solely from the slot, apparently it was okie dokie.

Why would you bother drawing the more than the bare minimum from the PCI bus if you're going to require a direct power supply connection anyway?

DELETE THIS!!!!1 NVIDIA IS PERFECT AND CAN DO NO WRONG!!!2 ONLY AMD CAN FUCK UP LIKE THIS!!!1

"In my understanding the 75 watt isnt the maximum limit, its just the default value on startup of the motherboard. The motherboard it self sets the maximum allowed watt per slot in the "Slot Capabilities Register" which you can configure up to over 300 watt per slot. In the bits 7 to 14 "Slot Power Limit Value" you can set 250, 275, 300 and above 300 watt. This will be multiplied with bits 15 to 16 "Slot Power Limit Scale" in steps x1 ,x0.1, x0.01 and x0.001. So its up to the motherboard manufacturer and the power management on it how many watt the slot is capable of. The Specifications do define the protocol and not the hardware specs of the PCI-E slot. If a manufacturer uses better parts which can handle higher amps on the contacts and the lines, they can allow the devie in the slot a higher power consumption than 75 watt via these registers. Sadly most people doesnt even read the specifications and judge things they dont understand."

this card was a massive fuck up in more than just this particular issue

>960
>two years old

>Budget level card with budget price performs at budget levels very well
Explain the failure.

3rd party with 8 pin with solve it.

cause nobody was stupid enough to own 960.
R-right?...

how do you get 1070's to crossfire?

as soon as i activate CF mode windows tells me my graphics card is not supported

weak bait

So 75W is the maximum a card can draw on startup because it is the minimum amount of power a PCI-e slot should be able to supply but then the card can negotiate a higher draw with the motherboard and it is up to the motherboard manufacturer to set the limits based on the motherboard construction and components.

Which basically means that the card was drawing over 150W because it the motherboard literally told it that it could.
Which means that tomshardware not doing a long overclocking benchmark for fear of damaging the motherboard was retarded because the motherboard basically informed the graphics card that it was able to supply that much power and if the motherboard died the only thing it would show us is that the motherboard was actually shit.

Source?

DELETE THIS

The 750ti did the same thing as well

>Some cards are using two 8-pin connectors, but this has not been standardized yet, therefore such cards must not carry the official PCI Express logo.
>literally all dual 8-pin cards are outside PCI-e specification
Oh wow.

jesus FUCKING CHRIST WHAT THE FUCK AM I SUPPOSED TO DO!

I don't want to buy a fucking 970 because >3.5 and because it is not ready for DX12. So I wait for this RX 480 from AMD, it's absolute shit and will KILL MY PC FROM POWER OVERDRAW. I can't buy a 1070 or 1080 until PROBABLY AUGUST.

WHAT THE FUCK AM I SUPPOSED TO DO Sup Forums!!!!!!!!!!!!!!!!! I just wanted to rebuild my PC int he beginning of the summer for a comfy gaming summer. Little did I know i would get bullshit thrown at me from all angles

It's actually much worse

oh wow just like 295x2

>Source: reddit
Thanks OP
Good post

It actually dosent matter

But which card is the % outside spec referring to?
>implying

>its a forum link
>PCI-E compliance require the circuit to provide a min 75watt to be certified

You can read further about the testing requirement, method etc over here by a company that makes tool to measure it
>teledynelecroy.com/doc/docview.aspx?id=7649

You are correct, however 75 watt isn't the maximum standard, but rather the lowest standard that a PCI-E circuit needs to deliver to be certified.

How high does it can handle literally depend on the motherboard itself.
Toms bench mobo obviously can handle at least twice the min specification.

How the heck such tech site freak out because of this?

pcper.com/reviews/Graphics-Cards/Power-Consumption-Concerns-Radeon-RX-480

>75 watt isn't the maximum standard

It is literally the maximum standard: en.wikipedia.org/wiki/PCI_Express#Power

The motherboard doesn't tell it anything. It might have monitoring/protection chip to cut off power to PCIe if it's too high (I doubt it does), but I know for a fact that motherboards can and do burn the 24 pin because of high PCIe draw in some multi card configurations (which is why a few high end boards have an aux power connector near the PCIe slots).
The power circuit on the PCIe slots should be dumb circuit. Just straight pipe 12V from PSU to the slot.

Any card worth anything should have 2x8pins or at least 8 pin and 6 pin.
I bet this beautiful fucker doesn't even draw on the PCIe slot.

lol nvdia sucks lmao

A motherboard is only built to go up to 75W on PCIE.

This is what happens when you go beyond that (from pcper article):

"I asked around our friends in the motherboard business for some feedback on this issue - is it something that users should be concerned about or are modern day motherboards built to handle this type of variance? One vendor told me directly that while spikes as high as 95 watts of power draw through the PCIE connection are tolerated without issue, sustained power draw at that kind of level would LIKELY CAUSE DAMAGE. The pins and connectors are the most likely failure points - he didn’t seem concerned about the traces on the board as they had enough copper in the power plane to withstand the current.

As we all know with hardware failures in PCs, this is something that could in theory happen during a single gaming session, or it might instead take months and months of gaming to wear down componentry. "

No, it isn't.
That specification standard you posted is out of date (2004?)

No it's not, it's literally the pcie 3.0 standard. 75W is the max allowed over pcie 3.0.

The problem is AMD for some reason decided to take shortcuts with the 480 and went far over this limit, causing all sorts of issues.

yes it is.

Do you have a citation that provides to the otherwise? Even if the report is ten years old he provided something authoritative and you have yet to provide anything other than posting "nuh-uh."

see

>According to the internet, the PCIE slot pulling over 75w was from 2 reviewers out of 20, of the other reviewers who were shared this information, non were able to recreate the scenario.
Stopped reading here. See

Do more than 2 reviewers even have the equipment to measure draw at slot?

It's difficult to measure, but for every single one that measured it, it has gone over the spec every single time.

Holy shit who is making these threads? You people don't know shit about electronics, the reference rx 480 is completely fine. It doesnt oc much but most clients don't care anyways about oc.
Oh and no it will not fry motherboards. I wonder what hive mind makes you think that...

Are you saying PCPer and TomsHard are unreliable or biased against AMD?

not everyone can afford it.

I want to like AMD, but it's really super annoying how fucking far you guys completely buy into AMD's hype time and time again.

"Nobody said it would perform as good as the things it doesn't perform as well as!"

Bull fucking shit. There were "980 TI PERFORMANCE FOR $200!!!!" threads plastered here all last week. Every single time it's the same shit here with AMD's stuff. Weeks (or even months) of hype and then when the thing drops and doesn't live up to expectations it's "nobody ever said that!" and then non-stop bitching. It's the same shit over and over again.

Or TPU?
>Normal gaming is higher than that limit, too. While nearly all motherboards and power supplies should be able to handle that, it still exceeds specifications, especially if you crank the power limit up while overclocking. Two 6-pin or one 8-pin would have maybe been the better power configuration.

Really? Is that why actual motherboard manufacturers are saying this?

"One vendor told me directly that while spikes as high as 95 watts of power draw through the PCIE connection are tolerated without issue, sustained power draw at that kind of level would LIKELY CAUSE DAMAGE. "

>it's really super annoying how fucking far you guys completely buy into AMD's hype time and time again.
Arguably AMD's worst factor is how much hype it spins out and how rabidly its supporters swallow it.

>There were "980 TI PERFORMANCE FOR $200!!!!" threads plastered here all last week.
Not to mention "THIS ISN'T A PAPER LAUCNH LIKE NVIDIA!"

Are there power consumption tests for PCIe draw done on other video cards? I'd like to see what is typical, not just based off specification

...

My fucking god, that is disgusting craftsmanship. They need to hire more than 11 year old schoolgirls to make their products. Cant even line shit up properly, and it needs TWO PCIE CONNECTORS WTF

Tom's does them all the time.

Tom's was actually the one who started this because they were the first ones to point out how far out of spec the 480 goes compared to all the other cards they've tested.

>There were "980 TI PERFORMANCE FOR $200!!!!" threads plastered here all last week.
I don't come here often, but the only thing I remember about the card was that its performance was to be in-between a 970 and a 980. Which is somewhat the case, even if slightly underwhelming.
But then again, this is Sup Forums, expect retards.

Nah, he's right. There were legit retards saying that. I asked them once sarcastically "what is wrong with having realistic expectations?"

They were saying it was better than a 980 and actually could reach a 1070 when overclocked.

Hilarious how so far off base they were.

There are still embellishments even now like

That thread is based off of claims made in a reddit post, which has been updated to read:

>EDIT: UPDATE FROM AMD:
>1) The RX 480 has passed PCIe compliance testing with PCI-SIG. This is not just our internal testing. I think that should be made very clear. Obviously there are a few GPUs exhibiting anomalous behavior, and we've been in touch with these reviewers for a few days to better understand their test configurations to see how this could be possible.

>2) Update #2 maybe the OP is confused. There is a difference between ASIC power, which is what ONLY THE GPU CONSUMES (110W), and total graphics power (TGP), which is what the entire graphics card uses (150W). There has been no change in the spec, so I would ask that incorrect information stop being disseminated as "fact."

>We will have more on this topic soon as we investigate, but it's worth reminding people that only a very small number of hundreds of RX 480 reviews worldwide encountered this issue. Clearly that makes it aberrant, rather than the rule, and we're working to get that number down to zero.


tl;dr OP's claim is factually inaccurate and OP is either an assmad nvidiot or a shill

The maximum PCI-E wattage is not enforced, its really up to the motherboard partner to allow how much a full lane can draw.
These compliance are nothing more but numbers on paper, and to get these certification they need to adorn the minimum standards which is 75watt/16lane.

SIG have never enforce these rules, they test it and give the certification to the vendor so that they can plaster the PCI-E shit on their box.

>75watt/16lane is a minimum voltage to be certified
>tons of motherboard with multiple slot/higher wattage per route are allowed to get these cert despite being 'over the spec'.
>SIG have no control over these manufacture and they can cram as much as they want as long as they meet the minimum spec SIG imposed to get the certification

In short, 75watt/16 lane is what SIG set, but not imposed on these manufacture.
Depending on the motherboards, going over the spec is completely safe as high as the motherboard allowed it to do/

>but mobo doesn't have control over the voltage

Wrong, in the most recent revision that I read, PCI-E offer flexible voltage control over the lanes through respective manufacture firmware.

You might want to read this, based on actual testing:

pcper.com/reviews/Graphics-Cards/Power-Consumption-Concerns-Radeon-RX-480

You are believing what AMD, who would never be biased or interested in defending their newest product, say over what independent reviewers are reporting?

The motherboard will of course have built in tolerances to go over 75W.

The problem is that it is not built for sustained use over 75W.

Straight from a mobo manufacturer on the issue:

"One vendor told me directly that while spikes as high as 95 watts of power draw through the PCIE connection are tolerated without issue, sustained power draw at that kind of level would LIKELY CAUSE DAMAGE. "

>there are a few GPUs exhibiting anomalous behavior

>You are believing what AMD, who would never be biased or interested in defending their newest product
>The RX 480 has passed PCIe compliance testing with PCI-SIG. This is not just our internal testing.
>This is not just our internal testing.


What the fuck is reading comprehension?

Why would over-spec power consumption fry the motherboard? Doesn't the motherboard itself regulate the amount of power delivered via PCIe bus?

It is not what AMD is saying. It is what PCI-SIG is saying. The guys that _make_ the fucking pcie spec greenlit it. Why would you need another opinion?

It's all marketing. The 750 ti drew more from the slot at load. Pcie 3.0 is able to handle 300w.

>What the fuck is reading comprehension?

Might have to ask you the same thing.

see

You know, funny thing is that the PCI-SIG fags might be lazy. I know that ecos consulting is. Those are the fags that give out 80+ gold ratings for PSUs. They don't even test the product sometimes and still give out a rating, because another product on the same platform already exists.

>What the fuck is reading comprehension?
Hmm. I wonder what is reading comprehension?
>This is not just our internal testing.
Said who? AMD? An independent reviewer? Under gaming loads or not? Overclocked or not? Of course AMD is not a corporation that would use underhanded tactics like posting half-truths. . .

>It is not what AMD is saying.
>UPDATE FROM AMD
If AMD is not saying it who is? Someone who hacked AMD?

It does. Motherboards will let you draw more than 300 watts through a single slot if you have a crossfire board. If you go out of bounds with the electricity the motherboard will just turn itself off.

Just a bunch nvidiots still but blasted for buying a 3.5gb meme card trying to report on something they have no clue about.

wait for AIB RX 480s with 2x PCIE connectors

Nah, the leaks were showing that it was a bit below a 980, overall, and people were speculating that it should overclock to give somewhere between the 980 and 980Ti in performance. That appears to be exactly the case.

It depends on the motherboards itself, as high as it allowed.
If the motherboard allowed the PCI-E to suck more than 200watt then the board is build to handle such load.

I am looking for it now, but I've seen some reviewer using low profile motherboard reporting a much respective power draw while having an issue overclocking it even with power limit set to high.
It is most likely that his motherboard is only allowing a lower maximum amount of watt drawn from it compared to these high end mobo these reviewers are using.

Also

>talking to VENDOR
>not directly to the manufacture/engineer themselves

why.jpeg

The slot can handle momentary loads of thousands of watts. The question is, how high is the sustained load? It's not that it can't handle the load, but the 12v contact points in either the 24 pin power connector or PCIe slot can get hot if you sustain high average power loading and that can damage or burn the contacts.

Spikes don't matter at all, but average kinda does. The killer is heat not voltage spikes. It's a dumb circuit. It doesn't have any delicate silicone that can be damaged by spikes.

>he believes they didn't actually do the tests or didn't pass PCI-SIG and are just lying about it

I hope nvidia is paying you well to act this dense.

Stop question Tom's Goyware.

>he believes they didn't actually do the tests
Straw man.

>are just lying about it
That's a possibility but I think it more plausible that the "testing" done was not representative of typical usage by a consumer, or as put previously, a "half-truth."

Sorry I think you got something mixed up here. Nvidia is the pathological liar company not AMD. Easy mistake to make :^)

>thousands of watts
>12v rail
>over 80 amps to reach 1kW

That's not going to happen. Ever.

>Pcie 3.0 is able to handle 300w.

Wow you really are dumb.

>2016
>not knowing tomshardshills is biased against AMD

>Why would you need another opinion?

Really?

Of course we want to see independent reviewers test it instead of blindly believing whatever AMD tells us.

And guess what...every single reviewer that has tested it has found the same power problem.

>Straw man.
Sorry, I miss quoted, that part was supposed to be in reference to: >Those are the fags that give out 80+ gold ratings for PSUs. They don't even test the product sometimes and still give out a rating

External tests have happened as well. Not all reviewers have had the same results.

>I believe in my guys and call the other guys liars despite the evidence
Whatever floats your boat. Unlike you, I'm not a fanboi who feels a need to defend or present one corporation over another.

Just like PCPer and TPU.
>I'm going to call everyone who doesn't align with my beliefs are biased