AMD: Speculation, Rumors, and Discussion (Archive)

Maybe is all what they can do without physically modify the card.

 

On the MoBo side its "5.5A MAX"

 

I'm sure if you think about it, you can figure it out.....

Only for voltage, which is not the problem...

 

You are correct. And they are exactly 4% above that in worst case.

 

Basically means amd released a card that is out of spec then.
90% engineers hired so how does one defend such big as mistake?

 

Umm... 6.2/5.5 = 1.127.. or 12.7%

 

I should have been more clear, I was talking about compatibility mode ON.

 

Overall I think the solution works.
But, why did they not also test with power target and/or OC?
It is as if they think no-one is going to play around with WattMan.
I am kinda surprised, and this is not about trying to catch out AMD, just to define what can be done with the reference card from a consumer perspective.

Cheers

 

Probably not enough resources.

 

Spec compliance wise it is exactly the same as before the "fix". I mean why bother releasing a "fix" if it doesn't actually change the situation....

 

Do only way this is 100% fixable is a recall and 8 pin. Simple as that.

 

Not really.... as long as they are within spec on the motherboard side they would only be liable for card/psu issues, which there shouldn't be any.

 

Only way is to drop frequency/voltage to get into the pci-e spec of the amps of the motherboard and everything else. They just don't want to drop its performance.....

 

Well obviously I expect reference design to kinda go EOL rather quickly.

 

To be slightly less cancerous? or toxic? "Now with 8.7% less Amps!"

 

It brought the PCIe slot portion to more or less within spec (some variance between different "quality" GPUs etc), and no-one really gives a crap about the PCIe power connectors, you can pull something.. was it near 200W from 6 pin without dangering the integrity of even lowest quality PCIe power connectors, even though the spec says 75W max, and many cards have pulled "over the spec" from them before, too.

 

the amps are still way over the pci-e spec. ~13% over

 

There is no "more or less". The current is either <= 5.5 Amps or it is not.

It is baffling really.... but what isn't these days.

 

and the excessive amperage draw was causing the vdroop if I'm not mistake.

 

May I suggest that the power draw alarmists are shunted off to their own thread?