Article on the RX 480 power issue.

I like the way they explain the issue. Personally I feel if the card is run at stock speeds there should be no issues. Also interesting is the quality of the VRM power delivery circuits on the RX 480 described in the article.

http://tinyurl.com/zd2u7kp

 

thanks.

i got some good read too.

people in 2016 bashing amd about not being "inside" of TDP... 160-170W instead of 150W TDP. when TDP never ever has had the meaning of a maximum power draw.

here all those cards go way over their indicated TDP:
http://ht4u.net/reviews/2009/power_consumption_graphics/index14.php

cards from BOTH venders actually exceeding TDP by up to 100 WATT
but i guess thats just before the time of where our main trolls where born, hence they just cant remember or see 10-20W increase as a sin.
"ahh but they promised 150W, how is this power efficient????"

and yeah the pice slot spec... we'll see about that. i guess they'll definitely use the weekend to find the best option to solve it (bios update , whatever)
more for calming down all that than because of it really causing any issues.
this pot-au-feu is unbearable by now.

 

PcPer's re-evaluation of the Asus GTX 960

http://www.pcper.com/reviews/Graphics-Cards/Power-Consumption-Concerns-Radeon-RX-480/Evaluating-ASUS-GTX-960-Strix

 

OK....That does not say the RX 480 is going to kill MB's....

 

No, just putting to rest the idea that others get away with it, so why not us. I think potential issues with MB's are currently being tested not only by review sites but also by AMD and the certification authority.

 

watch this -> 1080/1070/750/960/950 etc. all this GPUs exceed the PCIe specs...
He He maby they need unbias?
also read about 1080FE + blower and Hawaii 290x review + blower

Polaris - Power Problems or PCI Exaggeration?

 

undervolt and overclocking it. shrug

https://www.reddit.com/r/Amd/comments/4qtuu5/rx_480_undervolt_oc_1340mhz/
https://www.reddit.com/r/Amd/comments/4qupw4/super_psa_all_rx480_owners_please_attempt_to/

 

thats some new fodder in the whole matter! thanks...
will read that. looks promising.

 

OMG!!! All these people are blowing up their MBs!!!!!!

Seriously....this looks like someone at AMD messed up on the default voltages in the drivers and the next drivers will cure the issue. But AMD will still suck for some.

I bet AIB boards are going to be fantastic. Hope the premium over reference won't be too bad.

 

OC3D did talk about it. I do not know how can even be someone surprised after Fiji Power limit and voltage bump reducing OC performance below stock clock performance.

So, as they decreased clock and still got same results, card simply is entering empty cycles, doing nothing to stay within Power consumption limit over time.

I really wonder if people OCing GPU have that much different approach than when they OC CPU. In the end, your target is to get to highest clock with lowest stable voltage and throw Power target to sky.

 

You can also just undervolt it and stay stock speed(skip to 1:13 if you want to get to the point.):

 

Seriously tho... No effect on performance 28 watts less power used and core used 12 less... Now that would bring it's efficiency up a lot... It would be closer to that 2.8x amd said. But nah stock voltage on reference cards is just too high.

Lets say if the power draw of the card for hilbert was 166 and it lost 28-33 depending on source and dropped 138-133 watts it would be great but nah.

It is actually 15-20% more efficient with lower voltage.

 

vase please stop comparing total power consumption to TDP. It honestly shows you have no clue what you're talking about. There is obviously an issue with the cards, and judging by the comments released by AMD today, it seems to me like it's relative to the new system they implemented with the cards, that adjusts voltage automatically as time passes and calibrates for each individual card.

We'll see I guess. If any proper website had done a real architecture review, we would know more.

 

my post:

your post:

:question: :question: :question: :question: :3eyes:

so i say, TDP does not have the meaning of maximum power draw

and you say, vase please stop comparing total power consumption with TDP


i slowly but surely really doubt the comicality of all of this can be beat by anything i have ever witnessed in a debate.

 

TDP is NOT loosely connected to power consumption.


If "power consumption" is more precisely defined as maximum sustainable power consumption on real-world workloads, as it should,
then TDP essentially becomes "power consumption".
And you'd be wise when constructing a cooling solution to build the design capable of dissipating at least that much (heat/sec.)power.

Throwing out "real-world workloads" fence changes nothing. Again TDP=power consumption.
Everyone and their dog think that the two are not the same; they are.

 

Case in point: RX 480

It's power consumption is 162W.

How you should rate its TDP?
162W!

Because you want cooler capable of dissipating that much power.

The part that might not be all too obvious is that essentially ALL power consumption gets transferred to heat.

 

"The RX 480’s Elaborate 6+1 Phase Design Puts The GTX 1080 Founder’s Edition “High Quality” Components Claims To Shame"

:funny:

Temps
GTX1080 under load 82deg RX480 83deg
GTX1080 idle 33deg RX480 35deg

Noise:
GTX1080 under load 40deg, RX480 42dBa
GTX1080 idle 34dBa, RX480 37dBa


Look at the thermal image in the review from HH.
This +1 power phase close to the 6pin power connector is doing nothing.

http://www.guru3d.com/articles-pages/amd-radeon-r9-rx-480-8gb-review,7.html

 

no clue because of invoking TDP and total power consumption in the discussion,
when in fact 480 issue has nothing to do with TDP or with total power consumption:

 

No, TDP is the power level the card is designed for. It is just a target for the thermal design with a given cooling solution under realistic conditions and with a realistic load.

The actual Power draw can be higher resulting in higher temps or more noise from the cooling solution having to blow more air.

From the noise levels and temps in HH review it can be deduced that the ref card is already showing that little headroom is left.

And Hilbert tested with an open review setup.
In less optimal conditions like a closed case there might be no headroom left.

 

exactly, thank you noisiv!

also i posted the following quote here way before some random people starting to have (in their view) spotted an additional non-pcie-slot related power-issue as well (TDP, max draw, whatever they tried to imply shall be another issue in their eyes again)

http://forums.guru3d.com/showpost.php?p=5297746&postcount=22