AMD in Trouble? RX 480 Powergate

Meaker said:
Because where you see I you can replace it with V/R. The three are ALWAYS linked.

So your equation can be substituted with heat = (V/R)^2*R = (V^2)/R

Power goes up with the square of the voltage! So voltage really can't be ignored.

Basic electronics to use V=IR to substitute in values, it gets hammered into you even as you move further in so if we are harking on about it then it's because it's true ;)

While it's nice to assume voltage will remain constant, it never does!

Anyway I was being a bit pedantic I will admit, but as the draw goes up you will see the voltage drop. The contacts in this case wont be the limiting factor, it will be the trace widths and these tests are showing up those motherboard manufacturers who have been skimping on them to the PCI-E slot.

Still they are unlikely to burn out with a bit of extra draw if they are following proper engineering design to leave margins.
Click to expand...

Fair enough. I definitely agree with the last paragraph.
 
Rofl and now i've just seen Gibbo's post on OCUK forums about rx480 4gb being only produced in the first batch and that's it.

Man, the deception employed by AMD in this launch is making nV's worst look good. Maybe that nV chick was the one behind 970 fiasco too.
 
we all expected 4gb to not be available, this is a $240 card
 
pendragon1 said:
apparently its waaay higher than 180W! I just googled what a R9 M485X(for another thread) was and found this:

"We recommend the AMD Radeon R9 M485X for gaming with highest details at resolutions up to, and including, 1680x1050.
Being a mxm module card, the AMD Radeon R9 M485X does not require any additional power connector, its power draw is rated at 250 W maximum."

how the hell can xmx pull 250W without a problem but a regular desktop mobo cant? the power circuitry, traces and connectors in a laptop cant be that much more robust than a desktop board, can they?
Click to expand...

it's designed with that in mind, that said 250w is a lot, I doubt it actually pulls that much.
 
so amd drops a bios update that downclocks the inital versions of the card and locks the power down to prevent going over 70w from pci-e then releases a 8 pin equipped rx 480x with something else and offers free games or something to anyone who complains... it does depress me they fumbled so bad here....
 
Trimlock said:
Good video, never heard of him before.
Click to expand...

Neither have I, but he's the first to demonstrate one of the newest features being introduced with the AMD Radeon RX 480 ; PCI-E non-compliance(tm), a completely open source and unregulated alternative to electrical specifications
 
lolfail9001 said:
Neither card had nearly as much hype as RX480.

750 Ti violation as was stated about 750 Times already is not as severe as rx480's, not even close.

Also, i forgot, where was the exact source for 6990 violation of slot specification?
Click to expand...
LOL I would not bother clarifying Nvidia power consumption.
I was reading the comments at PCPer and Allyn must had responded over a dozen times so far stating it is the average consumption-distribution that one needs to consider mostly at Tom's Hardware and not Max, for reasons I have mentioned myself in the past that the max is instantaneous bursts, still useful to see the trait-characteristics for short dynamic behaviour but not much to do as a concern from overheating/failing PCI express/ATX12V/Molex components.
Sustained power is where it should be more focused, and as the chart is too scrunched (you really need the window widened to make better use of the chart with the spikes-spurs) one has to go by their numbers instead.
Where there is no number one needs to look for the yellow dotted line for average, he had to keep responding about the 960 as well lol.
Made me laugh anyway with how many times he kept responding on this :)
And the average when OCing is definitely a concern, or when 2x480 in a PC, or potentially when it is a budget build PC (this is where it needs also more validating).
Cheers
 
Last edited:
Dayaks said:
That's a really expensive board. Beyond me why someone would crossfire over a 1070 or 1080.
Click to expand...
Rabid pro-AMD anti-nVidia bias. The illogicality of nVidia/AMD fanboyism never ceases to astound me. "RX 480 at $200 is amazing! I'm going to wait for a $300 AIB card, though."
 
Dayaks said:
That's a really expensive board. Beyond me why someone would crossfire over a 1070 or 1080.
Click to expand...

Mostly AMD fan boys who were misled by the slides they saw before launch claiming the cards to overthrow the 1080 GTX.

These days AMD bois can even eat shit if it comes from AMD.
 
lolfail9001 said:
Rofl and now i've just seen Gibbo's post on OCUK forums about rx480 4gb being only produced in the first batch and that's it.

Man, the deception employed by AMD in this launch is making nV's worst look good. Maybe that nV chick was the one behind 970 fiasco too.
Click to expand...
I'm pretty sure that Finland didn't receive any 4 GB cards. At least two of the biggest shops got nothing. 4 GB cards aren't even listed. It does feel bit meh considering that GTX 970 is cheaper than the 8 GB version.
 
Tup3x said:
I'm pretty sure that Finland didn't receive any 4 GB cards. At least two of the biggest shops got nothing. 4 GB cards aren't even listed. It does feel bit meh considering that GTX 970 is cheaper than the 8 GB version.
Click to expand...

I'm sure RX 480 4GB was a paper launch just so they can claim it was $200. AMD fanboys still claiming it's a $200 card when all reviews and cards available for purchase are $240 8GB versions.
 
Tyns said:
I'm sure RX 480 4GB was a paper launch just so they can claim it was $200. AMD fanboys still claiming it's a $200 card when all reviews and cards available for purchase are $240 8GB versions.
Click to expand...


Lies, deceit and boat load of hype. Now the cards killing motherboards. What's next AMD?
 
pcgeekesq said:
It lo0oks like that the pictured 3-card system has not been turned on. Will anyone dare?

The motherboard looks to be an old ASUS Rampage IV (X79-based), maybe the Extreme edition? I guess there's no real harm if such an old board gets smoked ...
Click to expand...

1) I was right about the MB being a Rampage IV. Points to me. :)
2) Turning the system on (and running 3D mark) wound up damaging the on-board audio on the MB, according to "Cubeman's" reports.
 
Last edited:
Maybe the #betterred campaign refers to the color your PCI slot turns when gaming with the 480????
 
lolfail9001 said:
Neither card had nearly as much hype as RX480.

750 Ti violation as was stated about 750 Times already is not as severe as rx480's, not even close.

Also, i forgot, where was the exact source for 6990 violation of slot specification?
Click to expand...
The 750 Ti was pulling 64W average from the PCI-E slot. There was no "violation" there.
 
Yeah its peaks were higher, but peaks are ok as long as they are short.
 
FighterOH said:
So Razor, what du you make of this, i seen you at first said its an easy fix, but is there really, well you are probally one of thoose who know most on here, and as far as i can see, most of your posts latly are almost neutral :), so i ask.
Click to expand...


It is an easy fix for AMD, if they want to do it, that is up to them, because

a) it will make them look bad, because they have to accept the fact the f'ed up
b) they have to do one of two things either down clock the card or up the TDP with new power connectors and PCB changes And this will invalidate all current reviews
c) leave it as is and they get screwed latter on

Any one of these there is always a back lash of they screwed up. But B is the best way to go and the cheapest way to go.
 
Well, now we know why Raja wanted to jump right to Vega - even he knew what a pile this was.
 
Zion Halcyon said:
Well, now we know why Raja wanted to jump right to Vega - even he knew what a pile this was.
Click to expand...


Yeah they knew what they had, and right now with all the people finding out about it, it looks like they pushed it too far for a single six pin connector.
 
no i think they keep boost clocks depending on how hard the game is pushing it. we know not all games hit the card that hard.


For one reason I have no idea why AMD doesn't throttle in furmark. They are just stupid. LOL I just think they have their TDP limit messed up big time.

Look at nvidia, no way their cards breaks tdp under furmark. They will throttle the shit out of it. Now you got rx 480 trying to keep the got damn boost clocks no matter what lol. I think its the opposite of what nvidia does.
 
NKD said:
no i think they keep boost clocks depending on how hard the game is pushing it. we know not all games hit the card that hard.


For one reason I have no idea why AMD doesn't throttle in furmark. They are just stupid. LOL I just think they have their TDP limit messed up big time.

Look at nvidia, no way their cards breaks tdp under furmark. They will throttle the shit out of it. Now you got rx 480 trying to keep the got damn boost clocks no matter what lol. I think its the opposite of what nvidia does.
Click to expand...


I think most AMD cards do throttle in Furmark, not sure about the rx 480, but my 290x definitely did.
 
NKD said:
no i think they keep boost clocks depending on how hard the game is pushing it. we know not all games hit the card that hard.


For one reason I have no idea why AMD doesn't throttle in furmark. They are just stupid. LOL I just think they have their TDP limit messed up big time.

Look at nvidia, no way their cards breaks tdp under furmark. They will throttle the shit out of it. Now you got rx 480 trying to keep the got damn boost clocks no matter what lol. I think its the opposite of what nvidia does.
Click to expand...

Desperation is in the air. It's all lining up with what Kyle wrote. Using a 6 pin over an 8 pin is just mind-boggling - how much money is saved? Seems like a marketing move and it may cost them quite a bit.
 
Tyns said:
Desperation is in the air. It's all lining up with what Kyle wrote. Using a 6 pin over an 8 pin is just mind-boggling - how much money is saved? Seems like a marketing move and it may cost them quite a bit.
Click to expand...

There are other possibilities:

1) someone in engineering mis-predicted how much power the cards would draw, and AMD didn't have enough time to redo the PCB.

2) fab processes come in flavors: faster higher-power (such as a desktop processor might prefer) and slower lower-power (for laptops) for example. These flavors are produced by controlling things like doping profiles, you don't have to alter the chip design. So it is possible that the initial clock speeds on the lower-power flavor of the process, that they may have planned to use, wasn't anywhere near what AMD wanted, so they told the fabs to use the faster higher-power flavor instead. But they may not have had time (or may not have realized that they needed) to redesign the PCB for the higher-power parts.
 
You must log in or register to reply here.
Back
Top