http://www.tomshardware.com/reviews/amd-radeon-rx-480-polaris-10,4616-9.html
>We skipped long-term overclocking and overvolting tests, since the Radeon RX 480’s power consumption through the PCIe slot jumped to an average of 100W, peaking at 200W. We just didn’t want to do that to our test platform.
>Believe it or not, the situation gets even worse. AMD's Radeon RX 480 draws 90W through the motherboard’s PCIe slot during our stress test. This is a full 20 percent above the limit.
BLUNDER OF THE CENTURY
what the fuck were they thinking?
>>55336454
I guess it's too much for AMD to make a decent product. And after all that hype and promises. Fuck them
>>55336454
What do you expect from AMD, Every time there's something new from AMD, people jump on the bandwagon and say it's gonna be great. I really hoped this card would be good, I still use the 7970, which is think is one of the best cards since the 8800 from Nvidia.
>>55336454
Daily reminder Tom didn't have any problem recommending the 960 strix.
So if the 3rd party Rx Poo in Loo has more than 1 6 pin connector
Then it's safe?
I'm prepared to forgive the power issues because it's on par with a 980 Ti for $199.
>>55336509
not to mention there's a reason why some motherboard manufacturers put a dedicated molex power connector on the motherboard exclusively for pci slots
limits are limits but there's safety margins built in and everyone pushes the limit, it will be a non issue when non-reference cards come out with an 8pin or 2x6pin connectors
and because it's /g/
inb4:
>amdshills in full force
>housefire
>>55336556
Since when. it's worse than a 970.
>>55336572
It's on par with, at worst. Quit being dramatic, faggot
>>55336572
It's 15% better than a 970. 980 is 10% better than a 970 and 980 Ti is also 15% better than a 970.
>>55336542
Should be
>>55336454
I read an interesting theory.
According to another thread, AMD boosted the voltage to unsafe levels, so essentially the stock 480 is actually an overclocked 480.
They did this in order to make sure the card could compete with a 970.
This caused unsafe power usage, and got them into the situation we are seeing now, going outside of pcie spec.
>>55336625
It can't even beat a single 970. It's also beaten by amd's own last gen 390.
>>55336683
>It can't even beat a single 970
Stop bullshitting
>>55336683
fuck off nvidiot
>>55336694
>>55336734
Sorry the truth hurts
>>55336683
The 970 can't beat the 390 either.
R.I.P. AMD
R.I.P RX480
>https://www.reddit.com/r/Amd/comments/4qfwd4/rx480_fails_pcie_specification/
>>55336683
The RX480 is faster stop being such a faget.
Pic related (22 Games from the last 3 years). With more dx12 games coming out in the future the Rx480 is bound to become faster than the 980 ti. It's always like this. When the 780ti came out it smoked the 290X. Today the 390X smokes the much newer 980.
nVidia deprecates their chips when new gens come out and the 970 is the next victim to fall.
>>55336744
What truth? A $200 card is as good as a $350 card? It is a replacement for the 380 and 960.
>>55336776
Faggot confirmed
>>55336791
Interesting prices you got there. Meanwhile in the real world the 970 is only $250, while the 480 is sold out everywhere and only available on ebay for $350
>>55336799
Did I hit a nerv?
AMD BTFO.
>>55336806
>>55336799
You might want to read that thread, it's kinda important for the safety of your system if you get a 480
I bought a RX480, I am thinking about returning it.
What the fuck is this shit, there is abolutely no room for OCing anymore.
OC Gtx970 seems to be a better deal (still)
The amount of power it pulls from PCIe is beyond specs and could kill your motherboard.
I l-love my motherboard though. Fuck AMD.
>>55336568
+ just to add to this they don't even torture test their 1080 founders edition (instead take power data while playing metro last light), why should we expect to take tomshardware seriously if they can't even put out consistent benchmarks
hell, looking into this further it's fucking EVIDENT they show clear bias towards nvidia
from the rx 480 power usage page:
>The Radeon RX 480’s minimum GPU and memory clock rate is 300MHz, resulting in an idle power measurement of 16W (or 19W if you're using multiple monitors). That's simply too high for a modern graphics card.
however from the 970 power usage page:
>The GTX 980 WindForce OC gives us our baseline by idling just under 10 W, while the derivative 970 consumes almost 20 W.
I'm sorry, is the 970, a last gen card, not a modern enough graphics card?
think this is a one off occurrence? guess again
>AMD’s Radeon RX 480 draws an average of 164W, which exceeds the company's target TDP.
>The Radeon RX 480 significantly exceeds its official TDP.
the rx 480 has a TDP (not a power usage measurement) of 150w, that's 12% over the tdp, however the gtx 970 has a tdp of 145w(gtx 980 165w), yet when tomshardware reviewed the 970 guess how much power they got out of it, I'll give you a hint, it's not <200w, it's 239 fucking watts
what does tomshardware have to say on the issue?
>Nvidia's stated TDP becomes a distant dream. In fact, if you compare the GeForce GTX 980’s power consumption to an overclocked GeForce GTX Titan Black, there really aren’t any differences between them. This is further evidence supporting our assertion that the new graphics card’s increased efficiency is largely attributable to better load adjustment and matching.
they're fucking PRAISING it
(reference 980 hit 284w under their test, *72% difference* over official watt)
what a fucking joke
>>55336665
>I'm a massive faggot and need to learn to read
>>55336806
970 was 350 at launch retard
They had to put them all on fire sale two weeks ago when the 480 made them obsolete.
$200 is the msrp, and will be sold at that level when the unprecedented demand for the 480 tapers off.
They are really giving the 970 away at this point, it's irrelevant
>>55336454
So even though these sites get free hardware and massive dosh for reviews they decided not to. Why not have your mobo get fried and rake in the money from the amount of traffic your site would get from it.
>>55336831
Theory is that the stock 480 is in fact an OC 480, because AMD upped the voltage to drastic levels to compete with a 970.
That's why it can't overclock, because it already is.
That's also what is causing its power problems.
>>55336831
I bought one and I'm holding on to it for the time being.
If AMD's answer is to drop the clocks to reduce power draw through the PCI-e then yeah, I'll return it as well and go for an aftermarket one with extra pins.
>>55336845
You're only at risk if you overclock it.
Aftermarket cards should fix the issue, still pretty retarded on AMD's part though.
>>55336855
Actually the 970 started going down when the 1070 and 1080 were announced.
>>55336831
>>55336845
Quit lying same fags
It's literally a non issue.
See:>>55336509
>>55336864
>I bought one and I'm holding on to it for the time being.
I guess that is what I am doing.
I am waiting for AMD to address the problems.
>>55336872
>It's literally a non issue.
>PCIe average ~82 Watts
>PCIe max ~100 Watts
0/10
>>55336898
Now would be the time to show all the massive complaints 960 owners have if you want to convince anyone at all.
>>55336854
Please correct me if I'm wrong : it has a tdp of 150w.
75w coming from the pcie lane and 75 from the 6 pin connector ?
>>55336938
I am not recommending the GTX960, and never will.
Still the RX480 stock draws way too much over PCIe, which makes OCing impossible without damaging your mainboard in the long term.
I was also hyped, but these are problems which you cant just accept.
In the end I hope that AMD addresses these problems, otherwise the comparison between the RX480 Stock and GTX970 stock isnt fair.
They pushed it intentionally to beat the GTX970, yet the card is worse if you compare it to an OC GTX970 which would be more fair.
>>55336454
JUST
>>55336454
will ayymd ever recover
>>55336979
TDP is not power draw.
>>55336998
Well, then it draws 75w from the pcie and 75w from the 6 pin ?
>>55336979
AMD is actually trying to say it's a 110W card now, when the truth is it draws over 180W, far outside the TDP but more importantly far outside the motherboard PCIE spec.
>>55336454
AMD trying to achieve what nvidia's 480 didn't. True housefires.
>>55336963
>28nm vs 28nm
>Nvidia uses less power and produces more performance
>BUT TDP IS THE SAME AS POWER CONSUMPTION
Said nobody ever
Also
>480 consumes as much power as a 970 and has lower performance while being a 14nm card
>AMD tries to hide their power consumption by pretending it only needs a 6 pin connector
>Draws 200w from PCI-E that's rated 75W
LITERALLY JUST FUCK MY SHIT UP SENPAI.
I can't wait for peoples budget builds to start exploding
>>55337009
it draws 75w from 6 pin and up to 200w from pcie which is the problem
>>55336984
Plenty of people bought a 960 strx, none had problems. Still has 5 stars on newegg. It's funny only Tom's card had this issue.
And it's on par with the 970, you really need to get a better grip on reality.
>>55337031
>none had problems
Yet
>>55337049
>2 years later
>>55337026
>>55337022
How do reviewers calculate how much it draws from the pcie and from the 6 pin ?
I understand placing a meter between the 6 pin and the RX 480. However how do they calculate for the pcie slot ? Special mbo software ?
I also understand they can calculate by placing a different card and calculating total power draw but still doesn't answer the pcie power draw question
>>55337031
>It's funny only Tom's card had this issue.
The thing is, not only Tom reports these problems.
Many different benchmarks show it, also german ones.
>And it's on par with the 970, you really need to get a better grip on reality.
No not really, benchmarks show that the 970 is faster.
Also you are comparing a new card with a card which is older than a year.
In reality the 480 should be way better than the 970, even if it's on par it is not worth the money if you take in consideration that it draws so much watt from the PCIe slot.
That's pretty desperate.
>>55337080
Osciloscopes
>>55336766
True but the 970 beats the 480
Quite a few bugs with this gpu already then, mainly the frying cheap mobos souncards & the nasty stutter in games
>>55337067
>2 years later
That's only drawing an extra 33% from the rated 75 w.
The 480 is drawing 260% more @ 200w from a 75 w rated PCI-E slot.
>>55337092
Between the pcie slot the card ?
>>55336779
By the time dx12 is fully operational and enough games are out to warrant it, the 480 will be nothing but a doorstop. (its already halfway there)
Kyle was right.
>>55337100
Special riser port.
>>55337100
If you're reading what the pins are drawing. The rest of it has to come from the PCI-E numbnuts. It isn't a nuclear power plant making it's own electricity is it?
Why do people always feel the need to defend something which is obviously a failure?
First it was the "missing" 0,5GB ram from the GTX970, now it is the way to high PCIe average Watt.
God I hate shills.
AMD fucked up this time.
>>55337100
Power Consumption
Test Method Contact-free DC Measurement at PCIe Slot (Using a Riser Card)
Contact-free DC Measurement at External Auxiliary Power Supply Cable
Direct Voltage Measurement at Power Supply
Test Equipment 2 x Rohde & Schwarz HMO 3054, 500MHz Digital Multi-Channel Oscilloscope with Storage Function
4 x Rohde & Schwarz HZO50 Current Probe (1mA-30A, 100kHz, DC)
4 x Rohde & Schwarz HZ355 (10:1 Probes, 500MHz)
1 x Rohde & Schwarz HMC 8012 Digital Multimeter with Storage Function
literally says on the page
>>55337012
They never claimed that, period. Some other companies failed to understand AMD.
Another case of hype exploding.
JUST wait until an aftermarket board comes out that's designed properly or get a 1070 and leave the worries behind.
Under load it only uses 3w less than my 1080. what the actual fuck
can't you just overclock the pci slot so the power is no prolbem??
>>55337115
>>55337116
Gotcha. Thanks
>>55337136
They did the whole point of this was a budget vr card which uses less power than the 390s.
>>55337031
The 960 strix does not go out of spec
the 480 does
Pic related, well under 150W
>>55337120
amd shills are on damage control m8
>>55337082
You can see a whole thread of people having the issue here, it's widespread:
https://www.reddit.com/r/Amd/comments/4qfwd4/rx480_fails_pcie_specification/
>>55337145
>can't you just overclock the pci slot so the power is no prolbem??
>>55337097
Look again
>>55336509
The only real problem is going to be if people have power supplies with split 12v rails because the motherboard might not be supplied enough current on it's 12v line to supply more than the 75W spec for the PCI-e (though most boards have multiple PCI-e slots so you would hope it the PSUs would be designed to at least be able to provide ~300W on the 12v rail that is feeding the board) which would make the build unstable.
It's not going to blow up or anything. The spec exists mostly because you need to know what you have to be able to supply.
>>55337138
>just wait
>just wait
>just wait
>>55336509
That was a Strix issue and not a 960 issue you fucking mong
The problem is the fact that AMDs 480 R E F E R E N C E card is pulling that much power though the mobo.
>>55337174
that's neo-/g/ for you
Watch this video:
https://www.youtube.com/watch?v=S4cUPpr3Yi4
With this soundtrack:
https://www.youtube.com/watch?v=PIxd9CDaygI
>>55337189
The strix doesn't even have an issue, it was well under 150w, see >>55337159
>>55337187
Hang fire for the 490, its the 1080 killer dont you know
>>55337145
kill yourself
>>55337187
>just wait for aftermarket
>just wait for drivers
>just wait for stock
>just wait for msrp
>just wait for amd to get their shit together
>AMD's Radeon RX 480 draws 90W through the motherboard’s PCIe slot during our stress test
JESUS
>>55337194
>trolls trolling trolls
"neo" /g/, lol
Power spikes are unrelated to the problem retard.
>We skipped long-term overclocking and overvolting tests, since the Radeon RX 480’s power consumption through the PCIe slot jumped to an average of 100W, peaking at 200W. We just didn’t want to do that to our test platform.
This isn't about this:
To be clear, your motherboard isn't going to catch fire. But standards exist for a reason. All of the components around the PCIe slot and along the path from the slot to the 24-pin ATX connector will suffer from the peaks. And depending on your platform's design, audio problems may also materialize.
This is about how they didn't want to overclock because it was drawing 200W from the PCI-E
those values that are above the rated power draw aren't spikes, they're literally averaging 90w
>>55337227
>jokes on them I was merely pretending
>>55337159
>>55337200
That's average power usage, not maximum. The rx480 also averages at 150W, which means that it can peak to higher power usage during heavy stress, which means it uses more power than what it should.
So, when are aftermarket cards dropping?
>>55337233
You've gotta think that most people buying the 480 won't have mid to high end motherboards
This could actually cause damage to low end Motherboards
>>55337233
>To be clear, your motherboard isn't going to catch fire.
For people who cut costs by buying a cheaper motherboard, you know, the target market for this card, reality might be something entirely different.
>>55337262
>amd system with an oced 9590
>two of these cards in sli
>oced ripjaws ram
Some AMDrool do it now!
And film your PC.
I've not seen a mobo explode yet. Please.
You'll get views on youtube I promise. Probably enough to buy a new PC.
>>55337275
It might not catch fire, but it might damn well blow up your other components by causing a power surge.
>>55337309
>owning AMD CPU
I'd rather die
>>55337316
In short, the only way this AMD card blows away Nvidia is if someone installs it on an old nForce motherboard
>There are some retards on this board that think this graph
>>55336509
is in anyway similar to
>>55337233
JUST FUCKING LOOK AT HOW CLOSELY PACKED THE POWER DRAW IS. IT'S NOT SPIKING IT'S LITERALLY DRAWING ABOUT 100W CONSTANT
>>55337009
TDP stands for thermal design power which basically means how much watts of heat a cooling solution needs to dissipate to keep a component (such as a gpu die) within operating parameters.
JESUS CHRIST
>>55337343
>your GPU is drawing as much power as your monitor on idle
AMD engineering everyone.
>>55336454
https://m.reddit.com/r/Amd/comments/4qfwd4/rx480_fails_pcie_specification/?utm_source=mweb_redirect&compact=true
Would a MSI z97 PC mate motherboard be considered "low end"? I bought a RX 480 and I don't want to potato it.
>>55337326
it would be extremely painful
>>55337328
>the only way this AMD card blows away Nvidia is if someone installs it on an old nForce motherboard
Good one.
>>55337359
It should be fine but don't try overclocking it. Bad things will happen.
>>55337343
DELET THIS
>>55337359
You're gonna be fine. Just don't crank your OC settings all the way up. And also
>getting reference cards
>>55337359
I think the limting issue would probably be more with your PSU. Since the mobo doesn't actually regulate the PCIe power, it simply distributes it to all the mobo components that need it.
The thing is that the traces might be weak and get hot.
http://www.hardware.fr/articles/951-9/consommation-efficacite-energetique.html
>Hardware.fr confirms this too. They also have a retail version (Sapphire 480) which exhibits the same problem. They also confirms the power usage going over 150W with both a review and retail version of the card.
AYY
Y
Y
>>55337375
gotcha
>>55337377
integrated graphics have left me impatient.
>>55337383
got a evga supernova g2 750w, hows that famalam?
why not just have more 6pins or shit for the gpu?
>>55337480
because poo in loo wanted to decieve the press that it's power efficient
>>55336776
Holy shit I shitposted so much in AMD's favor but this is too much. The 480 is an utter disappointment. Pascal is shit and I hoped AMD could do something good. I'm so glad I bought Maxwell.
The 480 is at 970 numbers, but not always (something way above tho).
It is a nuclear furnace where the 970 is not. It draws same if not more power (and from the MoBo PCIe port, which is bad). It costs slightly less (not in EU tho) but the materials are cheap shit. It has more RAM but who the fuck cares on a 1080p/1440p card.
I wanted to love this card but AMD just let us down one more time. NVidia will be free to cuck us more now.
Let's hope Vega and Tesla will do something fucking good.
>>55337359
it's the cheapest entry level z97 board, I wouldn't risk it
>>55337494
>this is just the beginning
>Vega
>gpu draw 30W
>pci-e draw 300W
The future is bright!
>>55336454
>put furmark on
>complains it is sucking power
>>55336963
first, I'm not claiming that amd are some paradigm of sticking to specs, merely that tomshardware shitting over the rx 480 for using 12% more power than the stated tdp is laughable when they were praising the 980 for using 72% more power than the power usage as well as the 970 for using 64% more power than specs under torture tests
secondly, if you want to get nitpicky with the details, the 290x never had an official tdp figure, tomshardware mentioned it was "250w (claimed)" while other websites put it at 275/290/300w (unofficial) - anandtech even criticise the card for vague tdp figure
however the tdp at 250w (let's be generous) means that tomshardware peak power usage on a torture test of 295w (see pic, and below) means that the r9 290x used... 18% more power than a vague tdp figure never publicly released, compared to 970/980's 64% and 72% respectively
however tomshardware also had 2 results for gaming figures (not pictured) which are 185w and 218w (70c, 90% fan load) respectively which are far below stated tdp
which leads me to third, the reason your picture has upwards of 364 watts on the reference card (45% more power used than 250w tdp) is that it's comparing an overclocked 290x (to a stock 970) - no fucking shit overclocked cards are going to use more power, nobody compares overclocked results to spec tdp because spec means stock
so if you want to do it properly, for *gaming* results and not torture tests,
970: 145w spec, 168w result: -- 15%
980: 165w 200w -- 21%
290x: 250w (generous) 185w/218w/295w(torture) -- 74%/87% less than spec and 18% more than spec under torture test
rx 480: 150w 164w -- 12%
I honestly don't know if this is bait at this point but I'm too far committed, either way I'm out because the thread divulged into "970 is better! 480 is better!" and I have zero interest in buying either card
>>55337541
GTX 980 does not use 72% more power, stop lying you fucktard
>>55337517
The problem is where it's sucking the power, fucking retarded amdrone.
>>55337343
I don't understand why AMD's power consumption is so fucking bad. Like seriously why are they so incompetent?
>>55337541
I honestly don't think you comprehend what the problem is here.
First - the card is drawing more than they advertised. 90% more. They advertised 86W. The card draws 170W. Which is ... whatever, all hardware companies lie. They shoudl've just sidestepped it and said it's for the lower end models or some shit.
The real problem though is, it's drawing most of it's power from the wrong place. And probably it's doing that because they wanted to hide the power draw altogether - herpderp, our card only needs 6 pins.
Yeah right.
They should've made it a 10 pin and everything would've been fine.
>>55337262
Honestly it could cause damage to any motherboard because it's beyond what any pcie slot in any motherboard is built to handle
>>55337631
>"Oh shit Rajeet! Our new chip is underperforming."
>"Let's just crank up the voltage and overclock it, Pajeet!"
Another day in AMD R&D.
>>55337248
It literally says maximum right on the chart, can you not read?
>>55337631
The memory runs at 3d clocks.
>>55337665
Worked for hawaii.
>>55337682
Didn't work so well when they went full on 6-pin retard though.
>>55337516
>>55337240
That's the fucking joke. You paying attention to retards makes you the retard, retard.
>>55337329
strix is actually worse, strix draws more than 225W from 75W limited connector.
You fucking underage.
>>55337665
I think that is seriously what has happened.
Check this out: https://www.reddit.com/r/Amd/comments/4qk6db/psa_this_is_why_the_rx480_is_exceeding_its_tdp/
>>55337767
No it's not, spikes happen, constant 90w draw is terrible and will literally melt the port.
>>55337767
No it doesn't, strix stays well within spec
>>55337120
desu reaching 3.5gb memory use is impossible unless the game is badly written, or you are trying to play 4K on a 970 (WOW).
not to defend Nvidia, it's still bad marketing, 3GB or 3.5GB for that matter would have been perfectly fine.
>>55337779
If they don't want to issue a massive recall, they'll have to reduce clocks to around 850-900mHz. Either way AMD had a pretty FUBAR launch with this card.
>>55337786
if you take strix it will about 85ish Watts average
spikes also spike average numbers.
>>55337362
but you are a big shill
>>55337789
its 75w connector it draws 225w from, it's not the card draw as a whole.
it's 75w connector, the 6pin one.
>>55337802
Again that's not what the argument is about. The problem is that the 480 was drawing 200w from the PCI-E slot during torture tests.
who cares let the dullards buy this housefire
im waiting for vega and vega 2
by then volta will spank them and amd/nvidia will drop prices to compete we all win
>>55337830
No it wasn't, 200watts as a whole WHILE OCed
rx480 draws 150watts on average and it draws more than 75w from pci-e slot.
Stop being retarded.
>>55336799
>Sorry the truth hurts
>>55337821
It doesn't even get anywhere near 225W, see >>55337159
>>55337837
see
>>55336454
>We skipped long-term overclocking and overvolting tests, since the Radeon RX 480’s power consumption through the PCIe slot jumped to an average of 100W, peaking at 200W. We just didn’t want to do that to our test platform.
Buy a used 970, they said
http://www.hardwarecanucks.com/forum/hardware-canucks-reviews/72889-radeon-rx480-8gb-performance-review-24.html
>>55337856
THAT IS AVERAGE YOU MORON!!!
IT DRAWS 225W FROM 75W 6PIN CONNECTOR AND DRAWS MORE THAN 75W ON AVERAGE YOU MOOOROOON.
YOU MOOOOROOOOON
>>55337857
>We skipped long-term overclocking and overvolting tests
Oh, they didn't test it? Oh look, they are speaking out of their asses, just like always.
>>55337875
No that is not average, that is max, read the graph
>>55337837
>being this retarded and not reading the article where this is being mentioned in the first place
>Thinking anyone cares that it's drawing 20+ w over the 75w rated PCI-E.
It's the 200w draw over PCI-E while overclocked that everyone is worried about not the 90w at stock. The worst part is that everyone is pretending that the stock power consumption over the PCI-E is totally fine, it's already over the rating. Overclocking it is basically guaranteeing your mobos destruction.
>>55337888
They did test it, just not long term, because they were seriously worried about destroying their motherboards.
That's how bad this 480 power issue really is. A professional reviewer didn't want to keep testing it because he was worried it would destroy his hardware.
>>55337888
>they didn't test it
They did, it failed and they didn't want to destroy their system to prove that AMD sucks balls.
>>55337889
then they are shit at testing
>>55337309
>9590
where am i going to find a 2000 watt psu?
when will this thing be back in stock?
>>55337930
Doesn't change the fact that the card draws 200w or 260% over the 75w rated PCI-E when anyone attempts to overclock it.
>>55336779
Can you link sauce
>>55337930
You have a serious reading comprehension problem.
Look at the graph you just posted, it says "Total System Power Consumption"
TOTAL
SYSTEM
POWER
>>55337930
Guess what. None of those cards drew more power over the PCIe slot than specification allows for.
>>55337798
its kinda sad that AMD is so awful marketing their own products
on the day of release they uploaded like 20 videos to their youtube channel about the card, instead of uploading 1 video daily prior to release like any sane person
even the dumbest youtuber knows its better to upload daily than to do that
So this thing somehow gets more power than allowed from 6-pin+PCI-e
Right? This is bad for motherboard.
OK, so how was it allowed to be sold in the market? It must've passed some test, that regulates the market for things like these?
How did it pass that?
>>55337930
Yes they havent got a clue have they in comparison to someone such as yourself.
Guys is it true that pcie slot can give 150w?
>>55337998
>OK, so how was it allowed to be sold in the market? It must've passed some test, that regulates the market for things like these?
>How did it pass that?
The theory is that AMD used an early version of the board to pass certification.
Then they realized it wasn't beating the 970 in benchmarks, so they upped the voltage, overclocked it, and then ended up with this mess.
>ITT: reference cards are shit
The reference AMD card draws too much from PCIE slot. The reference Nvidia cards thermal throttle like a motherfucker and lose their boost clock completely in mere minutes.
This isn't fucking news, you shitheads. We've known reference cards are shit for a long time.
Holy shit shut the fuck up and buy a partner card.
>>55338009
Only if you put a 480 in it xD
>>55338017
>>55337998
>>55338039
AMD actually bankrupt and finished?
>>55338039
>making a facebook-tier image
Cool job man nice!!!
Q and A pci sign
Q: Does PCIe 3.0 enable greater power delivery to cards?
A: The PCIe Card Electromechanical (CEM) 3.0 specification consolidates all previous form factor power delivery specifications, including the 150W and the 300W specifications.
75w is pci e 2.0 Standart.
>>55338077
Nah I just stole it off another thread in all honesty.
>>55338029
but Nvidia cards don't destroy your hardware when you attempt to overclock it.
This is important news about the hyped AMD RX 480 that everyone should know. Please stop trying to bury it.
>>55338078
The 300W is for including 8 pin connectors, see this: https://en.wikipedia.org/wiki/PCI_Express#Power
75W is the maximum allowed through PCIE
If power draw really is a problem do you think we'll see third party cards with custom coolers and OCs coming with twin 6 or 8pin versions?
>>55338138
If you want to overclock safety it's literally a must.
>>55338138
Obviously. Even my GTX 980 has 2x8pins. Which the Nvidia reference version doesn't.
>>55337343
Wait, what the fuck. This was one of my gripes with the 390, but I was glad Fiji fixed it for the future generations. AMD what the fuck are you doing? What about high refresh monitors? Fiji was fine, but Nvidia went full turbo on them.
>>55337861
> DX12
Not nvidia shill, but come on now.
> 970
970 is prev-gen.
It is very good at DX11, and performs adequate in DX12. It can also utilize all the Nvidia technologies, Gamework, PhysX, and so on.
It is a very power efficient GPU that can drive any game on high resolution at 1080p. The fast vram amount (3.5gb) is MORE than enough, ever to be filled. It is impossible to reach that limit, unless you specifically try to do so - such as, with custom written programs.
> DX12.
Only available on the telemetry infested OS aka Windows 10 - and it works properly in only ONE major title (Ashes of the Singularity).
Before you call me tinfoil, regular 10 is just way beyond fucked up. Ads, messed up updates, it's one hell of a mess.
>>55338475
con't
>>55337861
> DX12 ...
DX12 is a fucking weird spec.
https://en.wikipedia.org/wiki/Feature_levels_in_Direct3D#Direct3D_12
Scroll below for the feature matrix part.
On top of that, developers need to do that extra mile to optimize the game for DX12.
Simply having a DX12 render will NOT result in any more frames per seconds.
(Actually, see Tomb Raider, it results in LOSS of performance.)
And, if they do that extra work - you can already utilize more cores with DX11.
Heck, CPU is only a bottleneck in a very few titles. Such as UE MMO games, where shitty developers rape the engine.
See TERA, Blade & Soul for example.
tl;dr: DX12 is good, promising, will be something in 2-3-4-5 years.
When titles will support it.
When Microsoft decides NOT to fuck people over with 10.
When Microsoft decides to release Windows 10 LTSB to the public, and let them have a proper operating system.
When cards will have proper suport.
>>55336454
Pretty early in the century to make that call yet. Give them time for more.
Coming from someone whos last 2 gpus have been amd this looks pretty fucking shit
> DX12
> not Vulkan
shiggdddddddddy
>>55338495
>CPU is only a bottleneck in a very few titles.
Stop posting before you make yourself look anymore stupid.
>>55336625
in AMD(R) Enterprise(tm) Benchmark(tm) maybe.
>>55338475
>>55338495
>>55337861
(As you can see on the Wiki article, only Intel Skylake supports ALL DirectX12 (current) features.)
When companies market (AMD and Nvidia both) their cards "GPU12 compatible", they just mean the card can run DX12 render, but cannot really benefit from all the new good stuff.
Btw, if titles were available, and DX12 would give more FPS, I would use it, sign me the fuck up.
But they dont. Ashes? It ran with +3FPS for me with my GTX880M and 780, well within margin of error.
Other games...
> Source engine (from HL2-CS:GO): Nope
> Overwatch: Nope
> War Thunder (Dagor engine): Nope
> Blade & Soul (recent free-2-play MMO): Nope
> World of Tanks (another popular f2p title): Nope
> LoL: Nope
> Black Desert Online: Nope
> Rainbow Six: Siege: Nope
> MGS V (Fox engine): Nope
https://en.wikipedia.org/wiki/List_of_games_with_DirectX_12_support
From this list, BF1 (Battlefield Niggers Edition) and Watch Dogs will be proper DX12 titles.
Maybe Deus Ex.
>>55338563
I have no reaction image for this level of stupidity.
>>55338548
Vulkan is shit, even DX11 beats it. At least in benchmarks.
It's a successor to OpenGL and it performs better than OpenGL. And on Linux/Android/Apple, you don't have DirectX, so it makes sense.
>>55338571
So wghat
if I don't have Skylake, but I have say - 900 series GPU from nVidiot, I can't get all the DX12 benefits?
nvidia shills literally on suicide alert hoping this actually becomes a huge blunder
>>55338600
Vulkan is like DX12. If you don't properly optimize it as a developer, you get shit performance. You might as well just stick to DX11 if you think low level apis are shit.
Bad performance is caused by developer failure not the api. In DX11 it's caused by the api, since developers have very little influence over the hardware.
>>55338602
>if I don't have Skylake, but I have say - 900 series GPU from nVidiot, I can't get all the DX12 benefits?
If the title uses the feature-set, indeed, you will be missing it on your 980, or 1080 for that matter. Not that AMD cards do any better when it comes to features, see the article.
>>55338624
Well they had like raw benchmarks for Vulkan, and it compared worse to DX11. As I said that is NOT an issue. It is an open-source competitor and a successor to OpenGL. And it does very well on that area, it's truly the next render in the Linux/Open Source/mobile space.
I am not dissing Vulkan or Khronos, it's great work. But remember, Microsoft is big. Bigger. Much much bigger. So they have money to pour into DX12.
>>55338105
Post links to any proof of an RX 480 causing motherboard failure.
See, nobody has even qualified that it's a problem. Everybody is taking information and jumping to the conclusion that fits their narrative best rather than analyzing the information to glean meaning.
I'm not saying that an overdrawn PCIE slot is a good thing, what I am saying is that it's going to be isolated to reference boards, which you shouldn't buy in the first place, and to overclocked cards, which you shouldn't do to reference cards.
Again, Nvidias reference cards are also shit. The GTX 960 overdraws also, and actually spikes higher. I should know; I own a 960 and it regularly overdraws my PSU and causes system crashes, even when the system is idling. I know it's the 960, because the problem started immediately after I upgraded from my GTX 560.
Is it causing motherboard damage? I don't know. Neither do you. Neither does Tom's hardware.
>>55339267
>it regularly overdraws my PSU
Only thing that means is that you have a shit PSU. If it is consistently drawing more power over the PCIe bus than the specification allows for, you have bigger problems.
>>55339331
It's an XFX 550W 80+ gold PSU that had no problems with my old GTX 560, which has higher TDP.
>>55339389
And just how old is that PSU? Just because it worked fine 3+ years ago doesn't mean it's fine now.
also
>XFX
>>55339331
>Only thing that means is that you have a shit PSU. If it is consistently drawing more power over the PCIe bus than the specification allows for, you have bigger problems.
Like what and why would you have problems in the first place?
>>55339389
TDP isn't power draw. 560 has worse efficiency which means that it produces more heat at the same power draw.
>>55339412
If you're a budget GPU buyer, chances are you bought a budget motherboard that'll happily blow out if you push the PCIe too hard.
>>55339422
And that's what I'm asking. Why would it blow our? You can probably power 2-3 GPUs on such cheap MB with no problems if they use around 75W over each PCI-E. So why would your MB fry if you use a little less power through 1 PCI-E slot? It makes no sense.
>>55339445
>Why can't I use detonator wires for high voltage power lines?!?!
>>55338670
It doesn't matter how much money they put into vulkan or dx12. It depends on the Dev using it whether it delivers performance or not. You can't magically optimize the api or driver.
Dx11 on the other hand depends on nvidia or Amd driver, whichever one you use. Thats the difference. Vendor optimisation is possible in Dx11, not in vulkan or dx12
So let me get this straight
>on par with 970/390
>similar price
>can't OC for shit unless you want a cooked GPU
>PCIE power draw could lose literal house fires
Are people SERIOUSLY defending this blunder?
>>55338571
warhammer total war dx12 released today, gave my 290 a nice bump from 55 to 75 fps.
>>55339534
You bet, people are either shilling or in hard denial mode over their preorders.
Also don't forget ~1.3v core on stock.
>>55339411
All the information yo needed to understand the situation was in my first post. I'm not an idiot. My hardware is not damaged, or third rate. The issue began immediately after I replaced my 560. I considered an RMA, but after doing some research, I found this is common behaviour for a 960.
>>55339412
TDP is not power draw, you are correct. Maximum power draw is often much higher than TDP. The point is that TDP is used to indicate power draw, regardless, and the 560 not only runs hotter, but also draws more power on average than the 960. The difference is that the 560 never had issues with spiking.
>>55339267
I also have a gtx 960 (msi armor 2x series). I'll timestamp if you all want. I built my first rig with a corsair cx 600w and the system fans would spin for a second and then stop - so I assumed it was a faulty unit. Returned it, replaced with the evga 600w and the system booted, but every once in a while, especially on boot, the system will come to a complete freeze or restart itself. Drivers have been clean installed - and the problem persists on multiple driver versions.
I had never considered it was the gpu before, since it was advertised as "consuming extremely little power".
>>55339648
Go back to your battlestation or desktop screenshot threads.
>>55338589
>*action*
Get the fuck out.
>>55339636
fud is nice, fud is life.
>>55339648
games are a perfectly valid use of computers and technology - fuck off. If I were to discuss the characters or gameplay elements in GTAV then yeah, it would be off topic. If I were to discuss hardware performance in GTAV, then it would be on-topic. These threads will never leave /g/. If you don't like it, find another website to discuss whatever the fuck it is that you consider on-topic.
>>55339648
/v/ is a console board. Their builds use sandy bridge at best.
>>55339648
How triggered can one retard be
>>55339636
>>55339687
But when Nvidia does it doesn't count.
When their products go above spec they go in gently, softly, under a soothing green ambient light. That would never do anyone any harm guise.
>>55339687
That's aftermarket brand, one case.
This is the basic reference card that's broken to shit.
>>55339742
Spec is averages not peak. Even a 450w psu can easily handle 550w peaks as long as average stays below 450.
>>55339757
What the fuck are you talking about? It's the exact same situation; one board design with a problem. Except in this case with AMD, it's the reference board, which are shit almost every time to begin with, whether it's AMD or Nvidia. AMD and Nvidia design GPU chips, not the boards those chips go on. Expecting them to do a good job with the board is like expecting General Motors to be experts at building and maintaining highway systems. It's not reasonable.
Nobody is recommending you buy a reference card. You should not buy an RX 480 right now. Likewise, you should not buy a 1070/1080 FE. Reference cards are shit, and have always been shit
>>55336454
All they had to do was put a fucking 8 pin. How dumb can they be?
Still hoping for monster custom cards, and subsequent honest competition from nvidia.
>>55339864
It's ASUS breaking a GPU.
This is AMD's own reference card breaking specifications.
It's not an aftermarket one, like ASUS STRIX or whatever.
>Nobody is recommending
It is no fucking excuse to sell broken shit to customers.
>>55336454
>yfw AMD housefire meme was real
>>55336861
It's because of the 6 pin and shit cooling. They should have revamped when it was hungrier than expected. But they wanted poo in loos with $15 psus to be able to use it.
Personally, I was waiting for non reference before it was even a twinkle in pajeets eye.
>>55337936
China.
So, this is it, right?
I'm trying to think of how AMD could recover from this, but I can't think of anything.
AMD is kill.
>>55339887
The theory is that the board was in production when they realized the GPU wasn't fast enough, so every Rx 480 out there right now is over clocked. It would explain why it's so hot, overdraws, and doesn't over clock well; the hardware is already at the limit of what it's capable of. It's strange that the card doesn't overdraw the PSU directly, though, since a 6-pin connector can deliver close to 200W as long as the cable itself is heavy enough that it won't melt. Almost every PSU out there can sustain 120W on 6-pin without any problems. PCIE spec leaves a lot of headroom in case of hardware failure so that things don't burst into flames before fail safes kick in.
>>55339653
I know it only proves i have the box, but I'm not uninstalling my card just to shitpost
>>55339742
>But when Nvidia does it doesn't count.
Nvidia didn't do it, though. Asus did.
>>55339961
>12:16 am
oops
>>55339901
Keeping in mind that AMD is an IC designer, not a board manufacturer, explain to me why it's OK for ASUS to "sell broken shit," but AMD's REFERENCE (as in the "this is how to get the GPU to work" board) should be expected to be any good? Consumers shouldn't fucking buy reference boards. AMD themselves don't even manufacture the majority of these boards, the partners do.
At some point, the consumer has to take some responsibility for being fucking retarded.
>>55339984
There are two anons in this thread who have problems with their non-ASUS GTX 960.
Do you own a 960?
>>55340038
Have I said at any point that it is okay for ASUS to sell broken shit? No. But it's ASUS, not nVidia making a broken reference card. It's different.
>Shouldn't buy
Stop excusing AMD and nVidia from shit reference card work.
>>55340038
lmao
This guy, excusing AMD which knows the GPU in and out for making a shitty PCB for it.
>>55339960
That's a dumb theory since ghetto rigged coolers on these reference models are letting them OC to 1400ish
>>55339949
>release 200$ 970 killer
Oh my.
>200$ version cannot be sold because of faulty design
Thank you Raja Pajeet Mohamed.
>the aftermarket cards are all 250$+ because they all have to redesign the power delivery completely
Wow.
>The 8GB model is now 300$
Pretty cheap good job AMD.
>Nvidia drops the price of the 970 right before the aftermarket 480s are released
RIP AMD.
And everything was alright in the world.
I honestly, very much hope AMD dies this year or next. Because that means a true competitor might finally show up when AMD is sold.
>>55340038
All 480 boards, as it stands, right now, draw more than the pci-e specs limit. Because it was designed that way. Tge fault of all cards rests on AMD. The one aftermarket 960 or whatever was the only 960 to do so, it was entirely Asus' fault, not nvidia, and it only pertained to a single aftermarket card type.
>>55340038
>Consumers shouldn't fucking buy reference boards.
Why is this okay?
>>55340038
>>At some point, the consumer has to take some responsibility for being fucking retarded.
The custom boards come at a price premium.
Wasn't this card supposed to be a 200$ wonder?
If I wanted a 300$ card I'd get a 390x and be done with it.
So can anyone clarify what a "low-end" motherboard is? I ordered a 480 yesterday that I plan to use with a Gigabyte GA-B150M-D3H DDR3 Micro ATX LGA1151 and a
EVGA 500W 80+ Bronze Certified psu.
Am I just going to have to return it?
how much did nvidia pay for that scaremongering?
>>55340313
AMD funded it themselves.
It's what you get when you release an absolute abortion of a product.
>>55340313
They paid in AMD victim complex bux.
>>55340369
>only a small number
Yeah. Strange how it's only the reviewers with proper testing equipment though.
>>55340286
That's a low end motherboard. Why would you need to return it? Lets say it burns your shitty motherboard which I doubt, do you feel sad about it? I wouldn't.
Can AMD ever do anything wrong in the eyes of their fanboys?
This card didn't deliver on it's promise of being 980 levels of performance.
It overclocks just as bad as fiji
Its hot, draws the same amount of power as a 1070 while being vastly inferior in terms of performance.
Now we also know it was badly designed and fries cheap motherboards.
It's funny how you idiots called kyle a shill while he must be pissing his pants laughing right now, literally everything he said was true
>>55340067
>shit reference work
You keep acting like this is news. Again, reference cards aren't supposed to be good. They're supposed to show the bare minimum needed to get the shit to work.
>>55340082
Yeah, and that's why I'm waiting until we get partner cards to decide what to buy. I think the 480 will be capable of more with proper power delivery and cooling, but we haven't seen a 1060 yet either.
>>55336454
>NDA till the last possible moment
>i expected it to be just meh
>i didn´t expect housefires
>>55336454
So this massive tech website that gets unlimited free parts didn't want to break one motherboard? Bunch of fucking fags.
>Aberrant issue less problematic than 3.5
>Nvidia shills still tearing the board apart over it
>It doesn't matter
>They'd be tearing the board apart with a perfect launch
>>55340525
>reference cards aren't supposed to be good
They also aren't supposed to burn your motherboard to a crisp.
This shit is unexcusable, stop defending it like a blind fanboy
>>55340550
maybe they meant their pcie pass-through measuring thing i think i read that it´s custommade
>>55340556
>your motherboard burning down is less problematic than having 3,5GB of Vram
You have some weird priorities
>>55340494
<- this was never an issue
(150w all on pci, but of course it's a 3 year old card so why should we care) (it's worth reminding that this one is also a budget card, so same shit mobo and psu FUD applies)
>>55340556
Why do you think it's less "problematic" than 3.5?
>>55340591
fake
>>55340591
but it's rated for 75w
the universe will explode if you draw more than that.
>>55340591
>all on PCI
LOL WHAT. where do you get this bullshit from
