Nvidia/RTG Powergate

Brackle

Brackle

Old Timer
Joined
Jun 19, 2003
Messages
8,568
Since this now seems to be an issue with both companies. What are peoples take on Both Nvidia and RTG Failing the PCI-e specs. Regardless of spikes or usage. Both companies make cards that go over the PCI-e spec.

AMD Radeon RX 480 8GB Power Consumption Results (RTG 480)

Power Consumption: Gaming - GeForce GTX 750 Ti Review: Maxwell Adds Performance Using Less Power (750ti Maxwell)

Tomshardware complains about RTG having the issue, but totally ignored the 750ti from doing it as well.

So thats shows biased from Tomshardware sure we can ignore that part.....

What I want to know since we know both companies do it. How do people really think?
 
I like my stuff to run within spec unless I tell it otherwise (ex. overclocking). Doesn't matter what company does what, if it's out of spec, I won't be buying it (not that the 750ti was ever on my radar to begin with).
 
I would think that the larger power fluctuations on the 750ti (and 960 not shown), which not only exceed max power specs at the peak but also violates the "maximum current slew rate" variation spec, would be more damaging to components like a motherboard and a PSU.

If component damage is the huge concern, I would have thought that the sub $150 PCIe-power-only option would be the more concerning part, landing in more builds with low quality motherboards and PSUs... Especially since its spec violations are potentially more damaging to these lower end components...

If the uproar is about the general violation of spec, I would think that violating 2 specs would be more damning. The 750ti even actively promoted itself as being PCIe powered, so you would think it would have had a focus on PCIe specs...
 
Last edited:
jamesgalb said:
I would think that the larger power fluctuations on the 750ti, which not only exceed max power specs at the peak but also violate the "maximum current slew rate" variation spec, would be more damaging to components like a motherboard and a PSU...

If component damage is the huge concern, I would have thought that the sub $150 PCIe-power-only option would be the more concerning part, landing in more builds with low quality components...

If the uproar is about the general violation of spec, I would think that violating 2 specs would be more damning, especially for a card that actively pomoted itself as being PCIe powered...
Click to expand...

The RX480 violates the slew rate limitation as well, on top of this, the sustained power while gaming on the 12V rail is 14% the rated maximum. Spikes above the maximum shoudn't happen, and they do ; 155W spikes on RX480

Additionally 82W sustained draw on 12v vs 71.3W maximum (with 8% voltage tolerance applied)

So both cards violate the specs, but the RX480 violates it far more seriously

01-GTX-750-Ti-Complete-Gaming-Loop-170-seconds.png


64 W average

141 max

RX480; 86W average, 155 max

and the 750Ti doesn't have a six pin connector, it's pci-e only

LOL

If you think a sustained load 14% above the MAXIMUM is the same as spikes above the maximum then there's really not much anyone can do for you other than hope you never go near their electronics :p
 
jamesgalb said:
I would think that the larger power fluctuations on the 750ti (and 960 not shown), which not only exceed max power specs at the peak but also violates the "maximum current slew rate" variation spec, would be more damaging to components like a motherboard and a PSU.

If component damage is the huge concern, I would have thought that the sub $150 PCIe-power-only option would be the more concerning part, landing in more builds with low quality motherboards and PSUs... Especially since its spec violations are potentially more damaging to these lower end components (higher peaks, larger fluctuations)...

If the uproar is about the general violation of spec, I would think that violating 2 specs would be more damning. The 750ti even actively promoted itself as being PCIe powered, so you would think it would have had a focus on PCIe specs...
Click to expand...

Araxie effectively answered you in your thread.

Araxie said:
the fact that you are assuming prove that you just want to guess what's good and what's not good in a power circuitry trace from a motherboard. sustained load out of specs produce overheat in the traces which are far pronounced on low-end motherboards with less PCB layers, the heat produce degradation of the traces and also will draw the power from other sources of the motherboard to keep up with the demand, that's where the most issues with bad audio, bad GPU signal, RAM troubles begins, think in a typical city highway when new everything it's flat and smooth but the higher traffic the faster degradation the asphalt will produce, in a similar way works the traces in a motherboard, the more and more sustained overcurrent loads will produce more damage, without take in consideration that the PCI-E slot itself will work hotter, and you must know, those issues with bent PCI-E slots, with loose ports are just aggravated, without take in consideration the extra strain putted in the motherboard logical controllers (yes, the chipset) even worse yet for the public aimed by this kind of GPU, it's the public that doesn't care of specs, doesn't care of power, just care about money and probably are using the cheapest possible motherboard. want to keep with the examples? keep thinking on the degradation suffered by a highway full transited of civil cars versus the same highway full of 10ton trucks. that's why specifications are made for to avoid issues, and the RX 480 go above 14% of the tolerance supported.
Click to expand...

But regardless, they both have issues, why would you bother with either one?? Also why the hell does it matter which violates the specs worse? Is this going to devolve into another AMD vs NV pissing contest?

EDIT: I should add, I'm genuinely curious if PCI-SIG can or will do anything about any of the violations, cause it's honestly not something I'm familiar with.
 
Oniigumo said:
Araxie effectively answered you in your thread.



But regardless, they both have issues, why would you bother with either one?? Also why the hell does it matter which violates the specs worse? Is this going to devolve into another AMD vs NV pissing contest?
Click to expand...

This thread exists purely for this reason, the RX480 launched yesterday, with review embargo lifting on the day of availability. Why bringing up 750ti, 9600gt (seriously, it was brought up) and asus strix 960 is relevant is beyond me... If they are all out of spec, so be it, they are all out of spec, it is no less damning for the rx480
 
Ieldra said:
This thread exists purely for this reason, the RX480 launched yesterday, with review embargo lifting on the day of availability. Why bringing up 750ti, 9600gt (seriously, it was brought up) and asus strix 960 is relevant is beyond me... If they are all out of spec, so be it, they are all out of spec, it is no less damning for the rx480
Click to expand...

So you do see the hypocrisy?
 
I don't think anyone here can really make a conclusion one way or another that either of the example cards is an issue - unless you design high powered PCI-E cards for a living.

There's a lot of folks acting like they are an expert in a field that they aren't.

People can quote, search, link, whatever they want.

Whether you think the cards are good, bad, one less bad then the other, etc, you're probably wrong.
 
Am I missing something?

According to Toms, the 480 is averaging 82W. The 750 Ti averages sub-75W (looks around 70W).
Looking at the ASUS Strix 960, it's averaging about 55W.

I don't see any other card averaging out-of-spec. Spiking above 75W isn't the same as staying there more than half the time.
 
TaintedSquirrel said:
Am I missing something?

According to Toms, the 480 is averaging 82W. The 750 Ti averages sub-75W (looks around 70W).
Looking at the ASUS Strix 960, it's averaging about 55W.

I don't see any card averaging out-of-spec.
Click to expand...

The 480 averages 82W on just the 12V rail. The maximum is 71.3W; 12V (+8% tolerance) @ 5.5A
 
TaintedSquirrel said:
Am I missing something?

According to Toms, the 480 is averaging 82W. The 750 Ti averages sub-75W (looks around 70W).
Looking at the ASUS Strix 960, it's averaging about 55W.

I don't see any card averaging out-of-spec. Spiking above 75W isn't the same as staying there more than half the time.
Click to expand...

So you know every card is out of spec?
 
TaintedSquirrel said:
Am I missing something?

According to Toms, the 480 is averaging 82W. The 750 Ti averages sub-75W (looks around 70W).
Looking at the ASUS Strix 960, it's averaging about 55W.

I don't see any card averaging out-of-spec.
Click to expand...
Technically, spikes like that violate PCI-E spec as well, the skews are way too extreme. But yes, some have come to think that "spikes" are now carrying same energy as averages.
 
Bowman15 said:
So you do see the hypocrisy?
Click to expand...
The hypocrisy? Bag on shitty reporting, that's the reason most people don't know about other cards violating specs. The only reason this is news is because of the hype surrounding the 480. I'd guess if the 750Ti was hyped to the same levels, people would have flown off the handle with plenty of forum posts to back it up. The same exact thing happened with the 970 (another hyped card) when it was found to be "3.5+.5", so there's no reason to act like this is a giant NV circle jerk smearing AMD.
 
Technically if you go over the maximum of 75 it's non-compliance, pure and simple.

Spikes are considered less serious because usually they are short and there's an increasing likelihood that there is some measurement error introduced considering resolution (500ns on some tests lol)

But seriously, for reference a GTX 1080 averages 40W with 62W spikes.
 
TaintedSquirrel said:
I'm not a lawyer so I could give a shit about violating the spec. Is it safe or not?
Click to expand...

Apparently PCPer spoike with motherboard people and above 90W is risky. So yes, if you overclock this thing, or buy those 1300mhz XFX cards, it is dangerous
 
TaintedSquirrel said:
Remember when we thought this card would be 100W or less? Apparently AMD thought so, too.
Click to expand...
By the way, Adored did get a card, didn't he? ROFL.

Also, computerbase.de actually made some fairly bold claims.

A) 110W is overall rx470 TBP, 80W is the ASIC power consumption (like 110W on rx480)

2) There is a 40 CU P10.
 
TaintedSquirrel said:
Remember when we thought this card would be 100W or less? Apparently AMD thought so, too.
Click to expand...

My lowest prediction was 120W, I was only off by 50W woop woop

lolfail9001 said:
By the way, Adored did get a card, didn't he? ROFL.

Also, computerbase.de actually made some fairly bold claims.

A) 110W is overall rx470 TBP, 80W is the ASIC power consumption (like 110W on rx480)

2) There is a 40 CU P10.
Click to expand...

That guy is gonna have a lot of spitshining to do particularly after the power fiasco ;)

btw, if you search "rx 480" on google, my reddit thread is one of the first hits
 
Bowman15 said:
You sure act like it..if you're not buying a 480 why give a fuck either way?
Click to expand...
Most people who are smart enough to be aware of this issue are smart enough to NOT buy ref models in the first place.
This is only a problem for idiots, masochists, and OEM's. Fuck those guys.
 
So, quick question. As a uneducated pleb when it comes to stuff like this, does someone more knowledgeable know if this could be fixed with a driver/vbios update?

EDIT: I'm assuming the quick and dirty method would just be to undervolt the card, but that'd seem like a shitty solution. Is that the only way, outside of partners adding connectors?
 
Oniigumo said:
So, quick question. As a uneducated pleb when it comes to stuff like this, does someone more knowledgeable know if this could be fixed with a driver/vbios update?
Click to expand...
Maxwell BIOS could control power distribution. I have to assume AMD's bios can do so as well.

I hope firmware can control it as well because having to employ a recall of first batch is the last thing AMD needs after the failure that is the GPU itself arch-wise. I am certain driver cannot (source: i have spent some time reviewing AMD GPU code in the past, most of functionality is offloaded to firmware).
 
TaintedSquirrel said:
Most people who are smart enough to be aware of this issue are smart enough to NOT buy ref models in the first place.
This is only a problem for idiots, masochists, and OEM's. Fuck those guys.
Click to expand...

They'll fix it, take care for those that already bought one and then life will move on. Then maybe you can as well and finally get some peace.
 
Bowman15 said:
They'll fix it, take care for those that already bought one and then life will move on. Then maybe you can as well and finally get some peace.
Click to expand...
Well, sure, after that we'll just a face a holy war about whether 1060 is better or it's worse. I, for one am certain of two things about 1060:

1) It will consume less power.

2) It will lose to rx480 in Hitman episode 1 and 3.
 
ir0nw0lf said:
:eek: Pretty detailed article there.
Click to expand...

indeed.. let's post some quotes for those lazy guys here..

This graph shows that result, running Metro: Last Light at 4K with the Radeon RX 480 at stock settings. The green line is the amperage being used by the +12V on the motherboard PCI Express connection and the blue represents the same over the 6-pin power connection. The motherboard is pulling more than 6.5A through the slot continuously during gaming and spikes over 7A a few times as well. That is a 27% delta in peak current draw from the PCI Express specification. The blue line for the 6-pin connection is just slightly lower.

At the top, you have a red and white line representing the voltage signal of the +12V rails from the motherboard PCIe slot and the 6-pin connection. Notice the drop on the white line of the motherboard +12V rail – during game play it is actually running at 11.5V while the PCIe 6-pin cable is a much safer 11.9V. When we exit the game at the 18:25:27 timestamp, power draw drops, current drops and the voltage returns to the same 11.9V we would expect to find. Looking back at the red line, the differences in power handling capability of the two sources become clear, as the 6-pin voltage barely flinches at the same current swings that caused 0.4V droop from the motherboard-supplied +12V source.

That voltage droop is caused by the current draw over the PCI Express connection, pins, and traces through the motherboard. Doing some quick math (0.5V drop at nearly 7A) tells us that the pins and traces are directly dissipating 3 watts of power in this state! What might be even worse for this voltage droop is that it affects all other PCI Express slots on our Rampage V Extreme motherboard when the primary slot was loaded to this degree. Any other add-in card that you run in the system with an RX 480 drawing this much power will be forced to run at the lower voltage. PCI Express does build in a tolerance level of +/- 8% for this value, so the rest of the system should remain stable, but one question would be what happens to that voltage when someone attempts quad-crossfire with overclocked RX 480s?
Click to expand...

Does any of this matter?

It seems clear at this point that the new AMD Radeon RX 480 does in fact draw more power through both the motherboard PCI Express connection and the 6-pin power connection than specifications state it should even when running at stock settings in certain gaming scenarios. The overdraw on the 6-pin cable is likely a non-issue; with power coming directly from the power supply and not passing through your motherboard and the fact that most cabling is built to handle higher power draw than we are seeing here, it’s very low on my list of concerns. The motherboard power draw is definitely something to keep an eye on though, especially given the voltage droop seen when motherboard traces are loaded to that degree.

The highest power draw I measured with the RX 480 at stock settings showed 80-85 watts of power draw at over 7A on the +12V line and 4.5-5.0 watts of power draw on the 3.3V line. These were consistent power draw numbers, not intermittent spikes, and users have a right to know how it works. When overclocked, we witnessed motherboard PCIe slot +12V power draw at 95+ watts!
Click to expand...

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.
Click to expand...

so yes it's something to be concerned said directly by the motherboard guys..
 
"Mine crashed my computer because of some Asus anti power surge thing... I think I'll leave OCing alone for a bit.. :/"


'let's wait until users report problems'
 
Ieldra said:
"Mine crashed my computer because of some Asus anti power surge thing... I think I'll leave OCing alone for a bit.. :/"


'let's wait until users report problems'
Click to expand...

LOL some asus anti power surge thing? WTF? LOL you know these people are just fucking tripping out now. PcPer literally kind of said there shouldnt be an issue at stock clocks. I call shits on this asus anthi surge thing lol. Now everyone that deosn't know what the fuck they are doing if they crash their computer they are going to blame on the card. I think most of these users are just making these things up. Crashed due to do some asus anti surge thing lol. Too funny.

You linked to PCperspective and what I got from that was this shouldn't be an issue correct? Although AMD is looking in to it.

To me it looks like he was trying to OC, I guess he didn't get the lesson. I have no idea why everyone is trying to OC this card when they know it doesn't OC well and we know power requirements are the issue.
 
NKD said:
LOL some asus anti power surge thing? WTF? LOL you know these people are just fucking tripping out now. PcPer literally kind of said there shouldnt be an issue at stock clocks. I call shits on this asus anthi surge thing lol. Now everyone that deosn't know what the fuck they are doing if they crash their card they are going to blame on the card. I think most of these users are just making these things up. Crashed due to do some asus anti surge thing lol. Too funny.

You linked to PCperspective and what I got from that was this shouldn't be an issue correct? Although AMD is looking in to it.
Click to expand...


You really misread the pcper article... I below e someone quoted them in this thread.

They say going over spec on 6pin is no biggie, but on mobo it's a big deal, incidentally, same thing I've been saying
 
Araxie said:
indeed.. let's post some quotes for those lazy guys here..







so yes it's something to be concerned said directly by the motherboard guys..
Click to expand...

mis information! he was talking about the 95w part that was when he was overclocking. during stock clocks it was 80. Stop trying to spread fud. He asked the motherboard guys about 95w being drawn when overclocking. Now why misquote him? Him saying it wasnt a cause for concern was on stock clocks. Gotta love people twisting words. Do people really have an agenda here?
 
lolfail9001 said:
Maxwell BIOS could control power distribution. I have to assume AMD's bios can do so as well.

I hope firmware can control it as well because having to employ a recall of first batch is the last thing AMD needs after the failure that is the GPU itself arch-wise. I am certain driver cannot (source: i have spent some time reviewing AMD GPU code in the past, most of functionality is offloaded to firmware).
Click to expand...

AMD had to address the Fury X for pump issues this time last year. They're no stranger to releasing beta cards and having the board partnerss deal with the RMA's.

Anandtech on Fury X Pump noise:
On Radeon R9 Fury X Pump Noise & Replacements

AMD for their part believes that the issue has been resolved going forward, and that is something that will need to be tested as the newer batches become available.

Finally, in the meantime there is the matter of the existing cards. AMD’s statement unfortunately doesn’t address the matter, though this is not unexpected since they aren’t directly selling the card to consumers. From what we’re hearing, the board partners are addressing this matter themselves, so any current Fury X owners experiencing pump whining would need to go to their board partner that sold their card for support.
Click to expand...
 
AMD played dirty this round. Trying to get extra juice to beat the 970 GTX. This card is a catastrophy so far. But AIBs won't suffer from the said issues.
 
Ieldra said:
You really misread the pcper article... I below e someone quoted them in this thread.

They say going over spec on 6pin is no biggie, but on mobo it's a big deal, incidentally, same thing I've been saying
Click to expand...

No I didn't He said at stock clocks it wasn't a cause for concern, only when overclocking. What did I mis read about that?
 
NKD said:
mis information! he was talking about the 95w part that was when he was overclocking. during stock clocks it was 80. Stop trying to spread fud. He asked the motherboard guys about 95w being drawn when overclocking. Now why misquote him? Him saying it wasnt a cause for concern was on stock clocks. Gotta love people twisting words. Do people really have an agenda here?
Click to expand...

read the article again please..
 
You must log in or register to reply here.
Back
Top