The 480: power consumption, PCI-E powerdraw

[SPBHM]

Not a good sign. First reports of damaged boards and/or pci-e slots.

https://community.amd.com/thread/202410

"ASRock 970 Extreme4"
far from a high end board, but that's not even one of the cheapest boards people are going to try using with an $200 card:thumbsdown:

 

[R0H1T]

Not a good sign. First reports of damaged boards and/or pci-e slots.

https://community.amd.com/thread/202410

Definitely bad news, let's hope it doesn't lead to any litigation in the US because that'll be a real nightmare!

 

[redzo]

Not a good sign. First reports of damaged boards and/or pci-e slots.

https://community.amd.com/thread/202410

It's a PR nightmare. I've got a feeling that'll escalate out of proportions.
"Dead motherboard?"
"The cause is uncertain to me but btw do you own a rx480?"

A motherboard designer should clarify this as I remember someone else mentioning.
Isn't the 75w max limit set for a good reason. What happens with those cheap motherboards designed to fit very close to this limit? You know, probably the rx480 will find its way mostly on cheap MB's since this is its purpose to begin with.

 

[richaron]

Not a good sign. First reports of damaged boards and/or pci-e slots.

https://community.amd.com/thread/202410

Overclocked CPU & GPU must have been pulling extreme power through that butdget mobo. People are too used to unbreakable hardware.

grumble grumble "back in my day" grumble "smoking CPU" grumble "ESD actually"....

 

[Eymar]

Not a good sign. First reports of damaged boards and/or pci-e slots.

https://community.amd.com/thread/202410

Well OP says trying to overclock to 1350. If issue with power on pci-e slot than pins on atx is probably burnt. I know had issue with OG Titan that required using pci-e power booster from EVGA from burning out ATX pins on X58 board.

 

[ioni]

Not a good sign. First reports of damaged boards and/or pci-e slots.

https://community.amd.com/thread/202410

A couple months ago, I had the same thing happen with my 7850. Except I didn't have another card to test the slot with. I was starting steamvr, pc froze, restarted, and it wouldn't start back up. The beep code said it was a video card problem. Tried rebooting several times, removing all power for a minute to try to "reset" everything, but always got the same beep codes. Switched to the other PCI slot and everything ran fine. When I got my 1070, I put it in the "bad" PCI slot (it's the only one it fits in unfortunately) and everything has been working great since.

So I'm still not sure what the problem was.

 

[antihelten]

but it's using the same power as a 270X
https://www.techpowerup.com/reviews/AMD/RX_480/22.html

it must be impossible to run a card like this just from PCIE?

it's well past 100W

TPU actually tested that exact model and it only pulls 61W, so it's just fine:

I'm guessing bios, the reviewers apparently got a special one that allowed them to swap between 4gb and 8gb modes (though seems like only one or two actually bothered to test it...)

Retail cards have been shown to have the same issue (Sapphire RX 480 drawing 166-170W), so it's not due to the special review BIOS.

 

[Azix]

so now all motherboard failures will be AMD fault

Though, I bought a used 7870 once that would not work in full x16 slots. Only worked in x8 slots and then it died completely. I think it died anyway. Maybe it needed a x4 slot

TPU actually tested that exact model and it only pulls 61W, so it's just fine:

https://tpucdn.com/reviews/MSI/GTX_750_Ti_Gaming/images/power_average.gif[/IMG

Retail cards have been shown to have the same issue ([URL="http://www.hardware.fr/articles/951-9/consommation-efficacite-energetique.html"]Sapphire RX 480 drawing 166-170W[/URL]), so it's not due to the special review BIOS.[/QUOTE]

average. it spends a good amount of time above the spec. but because of how averages work... average is low

 

[antihelten]

average. it spends a good amount of time above the spec. but because of how averages work... average is low

I know, averages has been what been discussed all along, since spikes are not really covered by the PCIe spec.

Someone linked to the average system power usage and claimed that the particular model of 750 Ti being tested must draw more than 100W based on this number, but as can be seen from TPUs numbers this is not the case, and thus the 750 Ti is infact running within PCIe specs.

 

[William Gaatjes]

Not a good sign. First reports of damaged boards and/or pci-e slots.

https://community.amd.com/thread/202410

I sure hope that it is not a lame attempt by someone with a grudge against AMD to be funny.

 

[Kenmitch]

Not a good sign. First reports of damaged boards and/or pci-e slots.

https://community.amd.com/thread/202410

1st alleged posting is more like it. Sucks for him if true.

You guys have to remember it's the internet where all things are possible.

For clarification I'm not defending AMD nor am I in denial. I'm just saying it's easy to say something on the Internet without providing any proof is all. I could just as easily say I bought a RX 480 and it fried my motherboard and be a iPad user that doesn't even own a desktop PC.

 

[zinfamous]

1st alleged posting is more like it. Sucks for him if true.

You guys have to remember it's the internet where all things are possible.

For clarification I'm not defending AMD nor am I in denial. I'm just saying it's easy to say something on the Internet without providing any proof is all. I could just as easily say I bought a RX 480 and it fried my motherboard and be a iPad user that doesn't even own a desktop PC.

What's next, you're going to tell me that people post fake reviews on Amazon and randos give poor ratings on Yelp simply out of spite?!

 

[someEEguy]

https://youtu.be/8Oc7zXhzlzU?t=388
https://youtu.be/8Oc7zXhzlzU?t=589

It seems like it could be that changing the power target in wattman may cause the pci-e to go out of spec...

 

[Abwx]

I sure hope that it is not a lame attempt by someone with a grudge against AMD to be funny.

That s BS since they dont even post a link to a pic, any phone can do shots of good enough quality.

For instance a good quality MB has no problem dealing with a few watt more at the PCIe..

But what about a bad/average quality MB..?.

Surprisingly it will have even less trouble dealing with such a delta, since it wont exist at all in this case, as with a more resistant routing in the MB most of the power will be deflected to the 6pin cable that will likely take 75%, if not more, of the card power..

Anyway the more i look at it the more i feel that it s just a question of agenda but certainly not a technicaly motivated point.

 

[amenx]

Not sure if anyone got beyond the first post, but a poster after him had same issue with an MSI board. A few others in other forums complaining about stability or audio issues too. As well as here, on old AM2 system:

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

I see somewhat of a common denominator, AMD based boards.

edit: I will add this to me amounts to very little. It can only mean something if these instances start increasing. So far not enough to draw conclusions.

 

[R0H1T]

Not sure if anyone got beyond the first post, but a poster after him had same issue with an MSI board. A few others in other forums complaining about stability or audio issues too. As well as here, on old AM2 system:

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

I see somewhat of a common denominator, AMD based boards.

edit: I will add this to me amounts to very little. It can only mean something if these instances start increasing. So far not enough to draw conclusions.

Wanna guess how many of them have cheap PSU's D:
The 480 is no doubt a catalyst in this but there's a fair chance that many of them systems were running on borrowed time, at least that's what IMHO

 

[antihelten]

The case you posted has nothing to do with CPU or whatever other component, it s a PSU failure apparently, so for the time you re just using some technical straw that is 100% irrelevant to the case discussed here..

GTX750 peaks at 140W and is most of the time well above the 75W line, so i dont see where would be the problem here, unless what is considered normal for one brand is not for another one...

You can't really tell how long the GPU was above 100W with the way Toms does their measurements and presents their graphs

PCPER probably explains this better than I could:

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.

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.

The fact that the average power measured by Toms is 64W pretty much also precludes the possibility of it spending the majority of the time above 75W, not to mention that the 1 second graph only shows it going above 75W roughly 25% of the time:

Compare this to the RX 480 which was measured to use 164W on average (with an average of 86W from the PCIe slot), making it almost certain that it spent the majority of the time above 75W plus the fact that it spends roughly 75% of the time above 75W in the 1000ms graph (it's a bit harder to tell here due to their new format, but it's there about):

So in other words, the RX 480 is treated differently to the 750 Ti because it clearly behaves differently (i.e. going above PCIe specs the majority of the time versus the 750 Ti that only goes above it the minority of the time).

 

[railven]

You can't really tell how long the GPU was above 100W with the way Toms does their measurements and presents their graphs

PCPER probably explains this better than I could:



The fact that the average power measured by Toms is 64W pretty much also precludes the possibility of it spending the majority of the time above 75W, not to mention that the 1 second graph only shows it going above 75W roughly 25% of the time:

<snip>

Tom's numbers for the RX 480 were far worse, so not sure why he even made that argument to begin with.

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.

200W peaks vs 143W and 100W average vs 64W.

 

[Abwx]

Tom's numbers for the RX 480 were far worse, so not sure why he even made that argument to begin with.

200W peaks vs 143W and 100W average vs 64W.

But the GTX960 peaks at 250W on the PCIe connector, why doesnt the site said that according to his own words :

power consumption through the PCIe slot jumped to an average of 100W, peaking at 200W. We just didn&#8217;t want to do that to our test platform.

So where was the outcry at the time, and why smaller numbers are suddenly an issue...?..

 

[96Firebird]

Average is what matters. Trying to tell others it doesn't changes nothing.

 

[Abwx]

Average is what matters. Trying to tell others it doesn't changes nothing.

Not at all, what create electrical shocks in connectors are sudden variation of the current, 250W peaks are more damaging than a 80W average that is anyway within the tolerance of the rated power, so much for the metric that matters for you, it may be good for whom has some kind of bias but not for whom talks of electrical laws...

 

[Rvenger]

I wonder if my black screen had anything to do with this over current into the PCI-E Slot. I was getting spikes to 1.275v and not even overclocking or raising the power target.

 

[IEC]

I wonder if my black screen had anything to do with this over current into the PCI-E Slot. I was getting spikes to 1.275v and not even overclocking or raising the power target.

Black screen on AMD cards is symptomatic for undervoltage conditions. I had a Sapphire Tri-X that initially worked fine but as it aged required a BIOS flash that gave +25mV to correct the black screen issue. Are you getting a black screen only on OS boot, or also in BIOS?

Overvoltages don't manifest themselves via black screens. Unless it's a permanent "black screen" like I got when I fried one of my cards via a voltage hard mod.

 

[96Firebird]

Not at all, what create electrical shocks in connectors are sudden variation of the current, 250W peaks are more damaging than a 80W average that is anyway within the tolerance of the rated power, so much for the metric that matters for you, it may be good for whom has some kind of bias but not for whom talks of electrical laws...

Anyone who claims another has a "bias" is not worth debating with, keep on with the childish comments and the good fight that this doesn't matter. :thumbsup:

 

[IEC]

Anyone who claims another has a "bias" is not worth debating with, keep on with the childish comments and the good fight that this doesn't matter. :thumbsup:

There's no need for ad-hominems.