Too much power to the people?
Does AMD have a problem on its hands with the power consumption of the new Radeon RX 480?
UPDATE (7/1/16): I have added a third page to this story that looks at the power consumption and power draw of the ASUS GeForce GTX 960 Strix card. This card was pointed out by many readers on our site and on reddit as having the same problem as the Radeon RX 480. As it turns out…not so much. Check it out!
UPDATE 2 (7/2/16): We have an official statement from AMD this morning.
As you know, we continuously tune our GPUs in order to maximize their performance within their given power envelopes and the speed of the memory interface, which in this case is an unprecedented 8Gbps for GDDR5. Recently, we identified select scenarios where the tuning of some RX 480 boards was not optimal. Fortunately, we can adjust the GPU's tuning via software in order to resolve this issue. We are already testing a driver that implements a fix, and we will provide an update to the community on our progress on Tuesday (July 5, 2016).
Honestly, that doesn't tell us much. And AMD appears to be deflecting slightly by using words like "some RX 480 boards". I don't believe this is limited to a subset of cards, or review samples only. AMD does indicate that the 8 Gbps memory on the 8GB variant might be partially to blame – which is an interesting correlation to test out later. The company does promise a fix for the problem via a driver update on Tuesday – we'll be sure to give that a test and see what changes are measured in both performance and in power consumption.
The launch of the AMD Radeon RX 480 has generally been considered a success. Our review of the new reference card shows impressive gains in architectural efficiency, improved positioning against NVIDIA’s competing parts in the same price range as well as VR-ready gaming performance starting at $199 for the 4GB model. AMD has every right to be proud of the new product and should have this lone position until the GeForce product line brings a Pascal card down into the same price category.
If you read carefully through my review, there was some interesting data that cropped up around the power consumption and delivery on the new RX 480. Looking at our power consumption numbers, measured directly from the card, not from the wall, it was using slightly more than the 150 watt TDP it was advertised as. This was done at 1920×1080 and tested in both Rise of the Tomb Raider and The Witcher 3.
When overclocked, the results were even higher, approaching the 200 watt mark in Rise of the Tomb Raider!
A portion of the review over at Tom’s Hardware produced similar results but detailed the power consumption from the motherboard PCI Express connection versus the power provided by the 6-pin PCIe power cable. There has been a considerable amount of discussion in the community about the amount of power the RX 480 draws through the motherboard, whether it is out of spec and what kind of impact it might have on the stability or life of the PC the RX 480 is installed in.
As it turns out, we have the ability to measure the exact same kind of data, albeit through a different method than Tom’s, and wanted to see if the result we saw broke down in the same way.
Our Testing Methods
This is a complex topic so it makes sense to detail the methodology of our advanced power testing capability up front.
How do we do it? Simple in theory but surprisingly difficult in practice, we are intercepting the power being sent through the PCI Express bus as well as the ATX power connectors before they go to the graphics card and are directly measuring power draw with a 10 kHz DAQ (data acquisition) device. A huge thanks goes to Allyn for getting the setup up and running. We built a PCI Express bridge that is tapped to measure both 12V and 3.3V power and built some Corsair power cables that measure the 12V coming through those as well.
The result is data that looks like this.
What you are looking at here is the power measured from the GTX 1080. From time 0 to time 8 seconds or so, the system is idle, from 8 seconds to about 18 seconds Steam is starting up the title. From 18-26 seconds the game is at the menus, we load the game from 26-39 seconds and then we play through our benchmark run after that.
There are four lines drawn in the graph, the 12V and 3.3V results are from the PCI Express bus interface, while the one labeled PCIE is from the PCIE power connection from the power supply to the card. We have the ability to measure two power inputs there but because the GTX 1080 only uses a single 8-pin connector, there is only one shown here. Finally, the blue line is labeled total and is simply that: a total of the other measurements to get combined power draw and usage by the graphics card in question.
From this we can see a couple of interesting data points. First, the idle power of the GTX 1080 Founders Edition is only about 7.5 watts. Second, under a gaming load of Rise of the Tomb Raider, the card is pulling about 165-170 watts on average, though there are plenty of intermittent, spikes. Keep in mind we are sampling the power at 1000/s so this kind of behavior is more or less expected.
Different games and applications impose different loads on the GPU and can cause it to draw drastically different power. Even if a game runs slowly, it may not be drawing maximum power from the card if a certain system on the GPU (memory, shaders, ROPs) is bottlenecking other systems.
One interesting note on our data compared to what Tom’s Hardware presents – we are using a second order low pass filter to smooth out the data to make it more readable and more indicative of how power draw is handled by the components on the PCB. Tom’s story reported “maximum” power draw at 300 watts for the RX 480 and while that is technically accurate, those figures represent instantaneous power draw. That is interesting data in some circumstances, and may actually indicate other potential issues with excessively noisy power circuitry, but to us, it makes more sense to sample data at a high rate (10 kHz) but to filter it and present it more readable way that better meshes with the continuous power delivery capabilities of the system.
Image source: E2E Texas Instruments
An example of instantaneous voltage spikes on power supply phase changes
Some gamers have expressed concern over that “maximum” power draw of 300 watts on the RX 480 that Tom’s Hardware reported. While that power measurement is technically accurate, it doesn’t represent the continuous power draw of the hardware. Instead, that measure is a result of a high frequency data acquisition system that may take a reading at the exact moment that a power phase on the card switches. Any DC switching power supply that is riding close to a certain power level is going to exceed that on the leading edges of phase switches for some minute amount of time. This is another reason why our low pass filter on power data can help represent real-world power consumption accurately. That doesn’t mean the spikes they measure are not a potential cause for concern, that’s just not what we are focused on with our testing.
Setting up the Specification
Understanding complex specifications like PCI Express can be difficult, even for those of us working on hardware evaluation every day. Doing some digging, we were able to find a table that breaks things down for us.
We are dealing with high power PCI Express devices so we are only directly concerned with the far right column of data. For a rated 75 watt PCI Express slot, power consumption and current draw is broken down into two categories: +12V and +3.3V. The +3.3V line has a voltage tolerance of +/- 9% (3.03V – 3.597V) and has a 3A maximum current draw. Taking the voltage at the nominal 3.3V level, that results in a maximum power draw of 9.9 watts.
The +12V rail has a tolerance of +/- 8% (11.04V – 12.96V) and a maximum current draw of 5.5A, resulting in peak +12V power draw of 66 watts. The total for both +12V and +3.3V rails is 75.9 watts but noting from footer 4 at the bottom of the graph, the total should never exceed 75 watts, with either rail not extending past their current draw maximums.
Diving into the current
Let’s take a look at the data generated through our power testing and step through the information, piece by piece, so we can all understand what is going on. The graphs built by LabVIEW SignalExpress have a habit of switching around the colors of data points, so pay attention to the keys for each image.
Rise of the Tomb Raider (1080p) power draw, RX 480, Click to Enlarge
This graph shows Rise of the Tomb Raider running at 1080p. The yellow line up top is the total combined power consumption (in watts) calculated by adding up the power (12V and 3.3V) from the motherboard PCIe slot and the 6-pin PCIe power cable (12V). The line is hovering right at 150 watts, though we definitely see some spiking above that to 160 watts with an odd hit above 165 watts.
There is a nearly even split between the power draw of the 6-pin power connector and the motherboard PCIe connection. The blue line shows slightly higher power draw of the PCIe power cable (which is forgivable, as PSU 6-pin and 8-pin supplies are generally over-built) while the white line is the wattage drawn from the motherboard directly.
Below that is the red line for 3.3V power (only around 4-5 watts generally) and the green line (not used, only when the GPU has two 6/8-pin power connections).
Rise of the Tomb Raider (1080p) power draw, RX 480, Click to Enlarge
In this shot, we are using the same data but zooming on a section towards the beginning. It is easier to see our power consumption results, with the highest spike on total power nearly reaching the 170-watt mark. Keep in mind this is NOT with any kind of overclocking applied – everything is running at stock here. The blue line hits 85 watts and the white line (motherboard power) hits nearly 80 watts. PCI Express specifications state that the +12V power output through a motherboard connection shouldn’t exceed 66 watts (actually it is based on current, more on that later). Clearly, the RX 480 is beyond the edge of these limits but not to a degree where we would be concerned.
The Witcher 3 (1080p) power draw, RX 480, Click to Enlarge
The second game I tested before the controversy blew up was The Witcher 3, and in my testing this was a bigger draw on power than Rise of the Tomb Raider. When playing the game at 1080p it was averaging 155+ watts towards the end of the benchmark run and spiking to nearly 165 watts in a couple of instances.
The Witcher 3 (1080p) power draw, RX 480, Click to Enlarge
Zooming in a bit on the data we get more detail on the individual power draw from the motherboard and the PCIe 6-pin cable. The white line of the MB +12V power is going over 75 watts, but not dramatically so, while the +3.3V power is hovering just under 5 watts, for a total of ~80 watts. Power over the 6-pin connector goes above 80 watts here as well.
Right now all the OEMs are
Right now all the OEMs are using AMDs reference design…. once they start creating their own PCBs and custom cooling this problem can be addressed and fixed. If it can’t already with software.
That would be a hope if their
That would be a hope if their design changes the way it works. If the AIB uses the ref card design and slaps their cooler on it, they could be looking at same problem as there is now well since its known they will have to pull those cards and redesign the card which will delay the launch.
right now they have ZERO
right now they have ZERO reason to use the ref design with another cooler.
Unless they love returns and CS working overtime.
Just feel sorry for people
Just feel sorry for people who have burnt out their motherboards.
If AMD had only used an 8-pin connector there wouldn’t be any issue.
They left no margin for electrical safety with the 6-pin connector.
Of course, AMD has no liability as technically you should not overdraw if you don’t overclock, but then you wouldn’t experience the same level of performance.
yea but to be honest the only
yea but to be honest the only motherboard fries are people with el-cheapo motherboards with suspect power supplies.
No solid motherboard, with a solid supply has failed.
the home grown foxcon mobo and greybox PS are the ones feeling the pain.
Leave it to the pros, not the bedroom geeks.
So as most of you know, I
So as most of you know, I work in Distro, and of the 100 480’s, nearly a THIRD have already been returned. Most of these are DOA, but quite a few are also doing in the field, and we can only imagine how many more will fail within then next 12-24 months. And you wonder why we do not ever suggest AMD CPU’s or GPU’s anymore. Utterly unacceptable failure rates so far, and it’s just the beggining. I called Raja directly to get an, “Were sorry, we’ll cover ALL expenses/loses that youre going to have due to this.”
But was told only,”The pricing already has the higher failure rates taken into consideration.” (So because something is priced a certain point, it should not work? Or what?) And was then told,”We really don’t have the budget to cover even $10K-20K, let alone a 40K loss on a newly released product”. No apology was given, and so my company very well might just drop AMD, again.
Rant over.
your company should have
your company should have never picked up the 480 from the start if they cared about quality and reliability.
Quality and Reliability are EARNED.. i have waited for the RX to EARN its way… its failing…why would I stock that product?
Do i like looking bad? Do I like returns?
MORONS
you cry about “utterly unacceptable” rates of cards that are installed in shit systems that get fried. wow…
so now you make up this bullshit story that failure rate was built into the model……………is that weed i smell?
or a shill?
both?
ok yea
you are a fucking moron working at a fucking moron company that will eat shit while it keeps making fucking moron choices.
drop amd by all means fucking moron
they can also drop your ass….fucking moron.
So as most of you know, I
So as most of you know, I work in Distro
say the guy named anonymous.
try to learn to troll, son!
So as most of you know, I
So as most of you know, I work in Distro
say the guy named anonymous.
try to learn to troll, son!
I’m posting this for Allyn .
I’m posting this for Allyn . GO NAVY bro! retired AE here. are you sure you weren’t doing more fire control stuffs LOL. Ryan.. stop letting you Bull dog take all the heat for you! hahaha. Love you guys and thanks for the contents, and try’n (I wants all the whiskey) to do the best you can for all. You’r an awesome crew. 9.99 lol
Thanks, brother!
Thanks, brother!
I have a Gigabyte ga-h110m-a
I have a Gigabyte ga-h110m-a Mobo, would this be ok to run it or is this Mobo considered old?
Sorry, after what I have read
Sorry, after what I have read so far I would NOT consider this gpu, as configured until the power management issue is resolved.
Thanks for a truly great
Thanks for a truly great explanation.
When you as a layman read the PCI EXPRESS BASE SPECIFICATION, REV. 3.0 you get very confused when you read…
”
Slot Power Limit Value – In combination with the Slot Power
Limit Scale value, specifies the upper limit on power supplied by
the slot (see Section 6.9) or by other means to the adapter.
Power limit (in Watts) is calculated by multiplying the value in
this field by the value in the Slot Power Limit Scale field except
when the Slot Power Limit Scale field equals 00b (1.0x) and Slot
Power Limit Value exceeds EFh, the following alternative
encodings are used:
F0h = 250 W Slot Power Limit
F1h = 275 W Slot Power Limit
F2h = 300 W Slot Power Limit
F3h to FFh = Reserved for Slot Power Limit values above
300 W
This register must be implemented if the Slot Implemented bit is
Set.
Writes to this register also cause the Port to send the
Set_Slot_Power_Limit Message.
”
Does the limitation here only regard what you can set trough programing and not the limitation of the hardware it self? It states ‘F2h = 300 W Slot Power Limit’ and slot means slot, right?
Here is the paper http://composter.com.ua/documents/PCI_Express_Base_Specification_Revision_3.0.pdf
Sorry, it also says “or by
Sorry, it also says “or by other means to the adapter.”
So in fact this information says noting regarding how much power you can pull from ONLY the PCI slot.
Do you think this has
Do you think this has anything to do with the physical location of the power phases ? They’re right above the power portion of the pcie connector, so the traces on the 480’s pcb are longer from the 6 pin, causing more resistance, which could have been counteredd with thicker gauge wires.
Also, I think the most
Also, I think the most obvious “solution” to this would be to lower the power limit, and then undervolt to reduce the power needed at each P states to maintain the same performance, but, if amd could have done that they would have, so we can expect that it wouldn’t work on all cards.
Also, I think the most
Also, I think the most obvious “solution” to this would be to lower the power limit, and then undervolt to reduce the power needed at each P states to maintain the same performance, but, if amd could have done that they would have, so we can expect that it wouldn’t work on all cards.
I’m probably missing
I’m probably missing something here, so sorry for being an idiot.
Chris Angelini @ Tom’s Hardware wrote:
The seventh phase supplies the memory modules with power using the PCIe slot’s 3.3V rail.
So ~5 Watts for the memory? Really? Memory controller maybe?
Edit: Maybe AUX Voltage? Some say the memory VRM is on the back of the card.
I saw a video from an LN2
I saw a video from an LN2 overclocker that said that the memory VRM was the one right next to the 6 pin connector, so unlikley to use 3.3V.
Yes, buildzoid has identified
Yes, buildzoid has identified the VRMs. They are on the back of the card and next to the connector just as you said. Maybe we watch the same LN2 overclocker 🙂
Maybe the 3.3V rail supply the VRM for the VRAM VDDQ (IO) voltage. That is probably a lot less power than VDD.
So is the power draw an issue
So is the power draw an issue ONLY when overclocking/increasing the power limit?
all the guy who dont support
all the guy who dont support amd are really masochist, we are in a dual word manufacturing scenario in both gpu and cpu, the role of amd is vital for all pc buyers, why some people cant understand this?
in this transiction moment could be very intresting to support (not necessary buy but write better word) amd, supporting amd is a way to decrease nvdia prince, how someone cant understand this?
No one is going to buy
No one is going to buy hardware they don’t want or need just because
I’ve build a high end system and I want a new card that goes well with it
So what’s my option?
Wait for half a year for Vega?
If history is anything to go by then all it will do is match a gtx 1080
But a month before Vega comes out Nvidia launches the ti version
Tsk
AMD needed a card that sells like crazy
This is not a good start
And it’s entirely AMD’s fault
Well, if the problem wasn’t
Well, if the problem wasn’t important then AMD would not be issuing a fix, simple as that.
Having an 8 pin connector may
Having an 8 pin connector may not be a fix for the RX 480 card. What troubles me is the exact same power draw from both the 75 watt 6 pin connector and the power draw from the PCI-E slot in the slides shown here.
Since the 6 pin connector comes directly from the PC power supply 12 volt power rails, it has no real barrier to provide much more power than the rated 75 watts. There is a sense signal pin but that is for the graphics card so that it is not fooled by plugging in a 6 pin power connector to an 8 pin socket on the card. Indeed, Toms bar chart shows the 12 line supplying 85 watts on average with a peak of 116 watts. So why doesn’t the very capable power cable supply the extra needed to prevent the high power draw on the PCI-E bus?
The only answer I can come up with is that there is not a “discrimination” device to unbalance the loads from the two sources. Could it be as simple as the design of the card has three VRMs hooked to the 12 volt power cable and 3 VRMs hooked to the PCI-E bus?
With the power droop bringing the voltage down to 11.5 volts for the PCI-E bus, it may be a simple test to trace out where this power goes on the graphics card by tracking what on the card has 11.5 volts. It may not be that simple (capacitors, drop down resistors, etc) but it is worth a try.
you just opened a can of
you just opened a can of worms when u said you will be testing it on X MB… now people are asking you to test on all kinds of MBs including old/new server MBs,OEMs,expensive,cheap,amd/intel … that kind of RND should have been done by the manufacturer >> anyways IF you are going to test old MBs just put in mind the degradation levels might vary depending on OCs how many cards it held how many sata ports where use you know the usual … again can of worms :S
Comprehension skills Ryan?
Comprehension skills Ryan? AMD promised an updated statement on Tuesday on progress, NOT an updated driver. Next you’ll run the “No promised driver fix” click bait headline..
Funny, I read the exact same
Funny, I read the exact same words as you did and I too thought that it said there would be a fix on Tuesday. Extremely nice job of verbiage by AMD as they left themselves a big out if they cannot come up with a quick solution.
Nice piece of work guys!!!!
Nice piece of work guys!!!!
Hey guys, i had the
Hey guys, i had the pleasure/displeasure (you decide) to watch buildzoid stream over on twitch just finished, stock-oc pretty much 83C+,watercooled stagnated at 31C flat all the testing got it up to a stable 1410MHz 50%power target. Then came the physical modding where he found out what is going on exactly with the PCB
https://www.twitch.tv/buildzoid/v/75850933 from around the 55min mark. Power draws were not measured due to lack of means
TL:DW the grounding on 6pin is not up to what it should be, half of the GPU is going from the PCIe 12v aswell without any limiter. This is a design flaw on the pcb itself aka hardware issue. This is in clear violation of industry standarts and potential safety-hazards. God knows how many cards are out in the wild, getting into systems that could face problems not in 2days time but in 3-6 months.
@Ryan Shrout & Allyn
@Ryan Shrout & Allyn Malventano, Everyone is talking about power draw of RX 480, interestingly there was an article on wccftech.com they underclock the RX 480 and say low power usage with no performance degrade. So i was hoping u guys can do a under clock article and how it affect Power Consumption Concerns.
Attention to RYAN…
The
Attention to RYAN…
The issue has been discovered
3 of the 6pin grnds are together, no center sense pin isolated – the top 3 vrms run to the 6 pin
the bottom 3 vrms run to the pci-e fingers
They are on separate circuits and not connected
Go to https://www.twitch.tv/buildzoid/v/75850933 54 minutes in to see how it’s out of spec
from what i read here and saw
from what i read here and saw with a multi-meter and this guy https://www.twitch.tv/buildzoid
Amd was drunk when designing the power side of the card. idk how a firmware will fix this its jacked to heck and back
Oh what a beatup. Worse than
Oh what a beatup. Worse than 970’s 3.5Gbgate which at least had a real issue at heart, we have had a number of cards and other hardware that exceed specifications over a long time. TDP’s themselves are a rubbery figure. If power draw through slots or 6-pins was hard limited we’d see almost every card unable to be overclocked
This is like the cops telling you that everybody died because you went 2kph over the speed limit. Total hyperbole
It’s funny how some don’t get
It’s funny how some don’t get it
No one cares this card or any other draws more power than advertised
It draws too much power where it should not, the weakest link in the chain
The PCI-E socket
That’s the only issue
And at least the GTX 970 only gave you worse fps
Didn’t actually damage hardware
And there’s still a lawsuit going on about that advertised 4gb