AMD in Trouble? RX 480 Powergate

Sheppard said:
Here is the issue Leldra.
You noticed a little thing, and you saw an opportunity to blow something up without quite understanding what you were talking about. You've made the argument that pulling a little extra power from a single PCI-E port isn't a big deal and that it becomes an even bigger deal when running crossfire. You then made the argument that the specification is for each slot. Ok if the specification is for each slot and it isn't compounding then it isn't a huge issue on two slots. You then made the argument that all slots power the power from common traces. ok so its now VERY evident that those traces are capable of providing much more than the value of a single slot because they would have to, to support crossfire correct? Some motherboards like mine have 4 slots and even more. Therefore there is no issue with the traces being able to provide enough power, even for crossfire, and the specification is for each slot and if the extra power on a single slot isn't a big deal then the extra power on another slot isn't a big deal because those slots have their own specification.
That being said a certifying body has passed the card and you are trying to argue against what that body has said therefore you must provide and present evidence of that argument you are presenting. It is not the job of the presentee to verify your argument for you by providing data/


IF the part used in the airplane caused a failure. A failure we have yet to verify occurring in any way in this case.
Click to expand...

I think I fully understand what I'm talking about thank you very much, I don't believe you have any intention of discovering the truth of this, you have already decided that this silly and blown out of proportion and that I have an agenda, and that's fine : ) I think many things about your motivations as well, but I do not think it worth the effort to communicate : D Who cares
 
he asked one question to AMD and they gave him a run around, he kept getting more and more information that pointed out the same problem, AMD no resposnses that are meaningful, the mods at Reddit delete his post. What more can ya ask for, its like a conspiracy theory in the making lol.
 
razor1 said:
he asked one question to AMD and they gave him a run around, he kept getting more and more information that pointed out the same problem, AMD no resposnses that are meaningful, the mods at Reddit delete his post. What more can ya ask for, its like a conspiracy theory in the making lol.
Click to expand...
Not necessarily. It is never that easy nor innocent. AMDs response was they have their engineering team on it now to see if they can duplicate the issue. And for now that is all we or anyone can ask. Now after a reasonable amount of time if they have no answer then by all means give the call to arms. But badgering or beating the poor dead horse will do nothing for an amicable response.
 
That I agree with it will take them time, but they need to close this screw up as fast as possible.

But I don't agree with TGP and all the other stuff they stated. That is beating around the bush and not talking about the problem directly.

Its the same thing with nV saying about the 3.5 and .5 partitioning of the vram on the 970 there is no excuse for not disclosing that even worse were the ROP amounts hence why the lawsuit is happening.

AMD is right now in a position to get ahead of this problem and its not a problem they can't fix quickly.
 
razor1 said:
So you would do something like that and wait for a plane to come down? Tell me what planes ya work on cause I'm not flying on them lol, J/K I am 100% sure you would never do that you are arguing for arguing sake, jokes aside, this is what an actuary is for........
Click to expand...
naw, the rules and regulations that I work under are so strict you cant fart incorrectly.
 
razor1 said:
That I agree with it will take them time, but they need to close this screw up as fast as possible.

But I don't agree with TGP and all the other stuff they stated. That is beating around the bush and not talking about the problem directly.
Click to expand...
As much as we would like full disclosure, I love information and as much as I can get, it is not ever wise for a company to directly comment on a subject in the heat of the moment. Now you are correct a quick response is best to avert any additional negativity but sometimes it is best to wait till tomorrow to speak on the issue.
 
JustReason said:
As much as we would like full disclosure, I love information and as much as I can get, it is not ever wise for a company to directly comment on a subject in the heat of the moment. Now you are correct a quick response is best to avert any additional negativity but sometimes it is best to wait till tomorrow to speak on the issue.
Click to expand...


Thats the thing, they did a quick response without thinking. "we are looking into it" is all they had to say. Instead they pulled out TGP 110 watts, and what not, WTF was all that stuff is what came to my mind when they stated that. Just trying to confuse the matter? I don't think that was ever mentioned before "TGP of 110 watts"
 
Also, how the fuck are TWO 480s more efficient than a single 1080. Someone really needs to bring that up again.


Cause that's the reason everyone and their mother expected 90W from this thing, stupid, I know.
 
razor1 said:
Thats the thing, they did a quick response without thinking. "we are looking into it" is all they had to say. Instead they pulled out TGP 110 watts, and what not, WTF was all that stuff is what came to my mind when they stated that. Just trying to confuse the matter? I don't think that was ever mentioned before "TGP of 110 watts"
Click to expand...

Typical AMD stats as always.
 
Sheppard said:
Here is the issue Leldra.
You noticed a little thing, and you saw an opportunity to blow something up without quite understanding what you were talking about. You've made the argument that pulling a little extra power from a single PCI-E port isn't a big deal and that it becomes an even bigger deal when running crossfire. You then made the argument that the specification is for each slot. Ok if the specification is for each slot and it isn't compounding then it isn't a huge issue on two slots. You then made the argument that all slots power the power from common traces. ok so its now VERY evident that those traces are capable of providing much more than the value of a single slot because they would have to, to support crossfire correct? Some motherboards like mine have 4 slots and even more. Therefore there is no issue with the traces being able to provide enough power, even for crossfire, and the specification is for each slot and if the extra power on a single slot isn't a big deal then the extra power on another slot isn't a big deal because those slots have their own specification.
That being said a certifying body has passed the card and you are trying to argue against what that body has said therefore you must provide and present evidence of that argument you are presenting. It is not the job of the presentee to verify your argument for you by providing data/


IF the part used in the airplane caused a failure. A failure we have yet to verify occurring in any way in this case.
Click to expand...

Look, it's all based on how the traces are set up. But to be honest a single trace can run to two slots. If that trace is rated for 150 Watts. 75 Watts per slot. Not a HUGE deal if one slot pulls a little more than 75 Watts as it will just pull more from the main trace if the trace and pin is rated for it. But that leaves LESS reserve for the secondary slot. Get it?

And it was reported that cross fire didn't overclock very well at all. This would indicate potential problems with power delivery and or heat as overclocking requires more power. They however had zero issues overclocking one to over 1300MHz. If they are pulling over 150 Watts in crossfire from the motherboard power bus STOCK, it's no wonder cross fire overclock fails hard.

As I said before these overruns are small BUT running out of spec could mean the following problems:
1. Overheating
2. Shorter component life
3. Excess noise
4. Instability due to heat/voltage sag/current sag

When I overclock I BUY quality components to help keep it in spec. (ASUS Sabertooth and Corsair AX850 power supply. I also pick a card with additional PCIe power connections.) Joe Smoe College Kid might have this system 6 years or more and he is using the cheapest components he can to build his first system like my nephew. I'm talking ASRock H170 boards and 550 Watt Bronze power supplies. I'm still going to recommend the RX480 to him provided they fix this F'up without kneecapping it's performance. Otherwise I'll tell him to wait for a 1060 if it performs similarly.
 
Last edited by a moderator:
If there is noticeable Voltage sag in these, then there is a VERY serious issue.

Keep your warranty information close by lol.
 
AMD's Radeon HD 6990: The New Single Card King

This is simple
At the end of the day as the PCI-SIG is a pro-compliance organization as opposed to being a standard-enforcement organization, there’s little to lose for AMD or their partners by not being compliant with the PCIe power specifications. By not having passed compliance testing the only “penalty” for AMD is that they cannot claim the 6990 is PCIe compliant; funny enough they can even use the PCIe logo (we’ve already seen a Sapphire 6990 box with it). So does PCIe compliance matter? For mainstream products PCIe compliance matters for the purposes of getting OEM sales; for everything else including niche products like the 6990, PCIe compliance does not matter.
Click to expand...
 
Trimlock said:
If there is noticeable Voltage sag in these, then there is a VERY serious issue.

Keep your warranty information close by lol.
Click to expand...

IF there is a problem
1. AMD will likely try to fix it in BIOS. Doesn't cost them a thing.

2. If the BIOS fix handicaps the card from initial reviews, and there is enough outcry, I bet AMD will be forced to physically fix the issue or offer a refund.

AMD could just cop out and go, "Driver and BIOS changes offer changes in performance to match a design envelope all the time" excuse. But that would piss off so many people.
 
DigitalGriffin said:
IF there is a problem
1. AMD will likely try to fix it in BIOS. Doesn't cost them a thing.

2. If the BIOS fix handicaps the card from initial reviews, and there is enough outcry, I bet AMD will be forced to physically fix the issue or offer a refund.
Click to expand...

Yea, usually Voltage sag, dips, dwells wtf you want to call them can be indicative of a hardware problem(resistance issue?). I only mentioned it because you did, I don't think these are causing shorts though.
 
Trimlock said:
Yea, usually Voltage sag, dips, dwells wtf you want to call them can be indicative of a hardware problem(resistance issue?). I only mentioned it because you did, I don't think these are causing shorts though.
Click to expand...

Voltage sag comes from running too much current through too thin a conductor
 
damage control; now claiming this is press bios only, implying there is a distinct press bios lol

at the same time, I have two reports of the same exact thing on retail cards, independent of course
 
I just want to make it known, if I wake up tomorrow and find myself being cited as a source by wccf or videocardz or something, perhaps giving me false credentials, I am going to end it. I will abstain from ever posting in a forum again, just so I can end it at that perfect moment of pure chaos and hilarity
 
NKD said:
Yea thats what I have been saying But Ieldra is on a roll, lol. They can just set a hard limit to 150w on reference cards. bump up the speed a little on the fan to retain boost clocks. LOL problem solved!
Click to expand...


You know that's not how thermodynamics works right?

The POWER CONSUMPTION is the problem here, not temperatures. You can't lower power consumption without LOWERING boost clocks, voltages, or reducing operating temperature.

The card sucks down 25w over-spec.

You want to lower leakage the 25w you need here, you're going to have to drop the temperature tens of degrees. That's not going to happen unless you crank the fan on this thing to 100%.

The AMD Radeon R9 Fury X Review: Aiming For the Top

Fiji claimed to save 20w by adding water cooling that cut temperatures by around 20C. To get a cheap blower to perform that minor miracle, you'd have to run at 100%.

So no, AMD will down-clock thje card, since I doubt a simple BIOS update will allow the to tweak something like power routing (from the slot to thje 6-pin). And if so, AMD is mis-representing this card's performance.

Unless of course you enjoy listening to a dustbuster the likes of which havn't been seen since the FX 5800 Ultra,
 
Last edited:
lol, its amazing how people have no freaking idea how this is a really serious issue, it's really concerning because isn't a thing that can be solved via driver or even a BIOS change, it's the power circuitry who decide how the card behave and from where it take the performance, so it would require a hardware fix to really solve this problem, a driver update to "prevent" this card consume that kind of power will insta-mean decrease in performance so AMD will be automatically called out for cheaters and/or unrealistic performance in reviews which its even worse as retail cards behave this way, changing the BIOS to limit the power tables of the card then will require to the end user flash the BIOS and not everyone will be willing to do that as it may brick the card, and you will also have a under-performer card so then again AMD will have a horde of people attacking them due the high amount of flash fails so insane amount of RMA will be present, If they don't solve anything then every single issue with any motherboard using an RX 480 will be assumed caused by the GPU which will not cover any warranty, then the horde of guys complaining will be even worse.. those cards will be avoided until AIB can supply cards with more than one 6pin connector.. the way the card decide to use more power from the PCI-E slot instead of the PCI-E power connector can't be changed easily..

It's a big fail no matter the side you decide to see.
 
Poor Rob, trying to clean up a mess he didn't make.
I'll hang on the sidelines until somebody figures out what's going on.
 
Ieldra said:
I heard someone here is buying four 480s for mining, perfect for cooking popcorn on the motherboard pcb
Click to expand...

powered raisers are the only savior for those scenarios.. (y) and probably, they will be in safer scenario than the average joe running 2 cards in Xfire while gaming with a low-end motherboard.
 
Araxie said:
lol, its amazing how people have no freaking idea how this is a really serious issue, it's really concerning because isn't a thing that can be solved via driver or even a BIOS change, it's the power circuitry who decide how the card behave and from where it take the performance, so it would require a hardware fix to really solve this problem, a driver update to "prevent" this card consume that kind of power will insta-mean decrease in performance so AMD will be automatically called out for cheaters and/or unrealistic performance in reviews which its even worse as retail cards behave this way, changing the BIOS to limit the power tables of the card then will require to the end user flash the BIOS and not everyone will be willing to do that as it may brick the card, and you will also have a under-performer card so then again AMD will have a horde of people attacking them due the high amount of flash fails so insane amount of RMA will be present, If they don't solve anything then every single issue with any motherboard using an RX 480 will be assumed caused by the GPU which will not cover any warranty, then the horde of guys complaining will be even worse.. those cards will be avoided until AIB can supply cards with more than one 6pin connector.. the way the card decide to use more power from the PCI-E slot instead of the PCI-E power connector can't be changed easily..

It's a big fail no matter the side you decide to see.
Click to expand...

If it's an issue with the RX 480 PCB design, all cards would exhibit this issue. They aren't as PCPer demonstrates. It would have also never passed SIG certification, which it did. If it's a limited number of cards, it's a flaw with a certain batch of cards, either bios or manufacturing.
 
Yeah, the whole point fo this complaining is to get people with the cards and equipment to take a closer look at this.
 
defaultluser said:
Yeah, the whole point fo this complaining is to get people with the cards and equipment to take a closer look at this.
Click to expand...

You just get people who can't test it and likely don't have a 480 to "take a closer look at this". What you need are websites that know how to test this to take a closer look at this.
 
variant said:
You just get people who can't test it and likely don't have a 480 to "take a closer look at this". What you need are websites that know how to test this to take a closer look at this.
Click to expand...
We had several already including Tom's hardware, definitely no lite weight there. But, i do agree the more people can replicate this with store bought cards the better.
 
We all overclock on ECS boards because of their highest quality so I don't see the issue. I'm sure ECS motherboards can easily handle 1500 watts per slot!!

This sounds almost as bad as the 3.5 gigs of RAM issue on the 970's. When all was said and done it didn't make any difference.
 
Presbytier said:
We had several already including Tom's hardware, definitely no lite weight there. But, i do agree the more people can replicate this with store bought cards the better.
Click to expand...
I just posted a PCPer test that doesn't exhibit the overdraw problems.
 
defaultluser said:
You know that's not how thermodynamics works right?

The POWER CONSUMPTION is the problem here, not temperatures. You can't lower power consumption without LOWERING boost clocks, voltages, or reducing operating temperature.

The card sucks down 25w over-spec.

You want to lower leakage the 25w you need here, you're going to have to drop the temperature tens of degrees. That's not going to happen unless you crank the fan on this thing to 100%.

The AMD Radeon R9 Fury X Review: Aiming For the Top

Fiji claimed to save 20w by adding water cooling that cut temperatures by around 20C. To get a cheap blower to perform that minor miracle, you'd have to run at 100%.

So no, AMD will down-clock thje card, since I doubt a simple BIOS update will allow the to tweak something like power routing (from the slot to thje 6-pin). And if so, AMD is mis-representing this card's performance.

Unless of course you enjoy listening to a dustbuster the likes of which havn't been seen since the FX 5800 Ultra,
Click to expand...

I think it's possible. On my 970 you can specify the default/upper limit per rail.

3_zps4kqfi4pd.jpg


4_zpsco6rinrk.jpg


Whether Polaris BIOS has this or not, remains to be seen.
 
You must log in or register to reply here.
Back
Top