AMD Will Address RX 480 Power Consumption With Radeon Software Update This Week

[kcarbotte]

AMD spent the holiday weekend working on a driver update that will fix the power draw concerns of the RX 480. The company announced that the update would be released “in the next 48 hours.”

AMD Will Address RX 480 Power Consumption With Radeon Software Update This Week : Read more

 

[jasonkaler]

Nice feedback and response time from AMD about addressing the issue.
Nvidia, take note!

 

[monsta]

AMD would not have had to provide you with a response and feedback if they had only not gimped their card

 

[EvilMonk]

MONSTA remember the nVidia GTX 970 / 3.5Gb episode? don't start shit it happens with both of them...

 

[Calculatron]

Cheers!

 

[Lord_Sunday123]

On the danger scale, 3.5/.5 Gb as a little less dangerous than drawing too much power. I'm not saying it is frying mobos, but it remains a possibility. I don't quite understand why either the 480 or the 970 were able to make it out of the factory without people figuring out that there were problems, but you'd think things like that would be easy to figure out.

 

[Vladraconis]

In the case of the RX 480, it's the old / cheap mobo's that are having issues with the extra amp draw. Maybe they just didn't think it was necessary to test the card on older / cheap mobo's. Which is a bit weird, considering this is a 200$ card, so aimed rather at those who don't have the $$ for more expensive components.

 

[nutjob2]

Apparently AMD's little bug will cause your motherboard to go up in flames, but overclocking everything from your CPU to your toilet seat is perfectly ok.

 

[Quixit]

No, overclocking is something you do knowing you could be destroying your hardware. When you buy hardware you expect that it will operate in spec by default. They're not even slightly related.

 

[nutjob2]

No, overclocking is something you do knowing you could be destroying your hardware. When you buy hardware you expect that it will operate in spec by default. They're not even slightly related.

Give me a break, AMD's pushing a power line won't destroy or damage anything. I'm sure AMD's card will operate in spec, it's a new card with bugs, but apparently only AMD gets crucified for that, everyone else gets a pass.

 

[kcarbotte]

No, overclocking is something you do knowing you could be destroying your hardware. When you buy hardware you expect that it will operate in spec by default. They're not even slightly related.

Give me a break, AMD's pushing a power line won't destroy or damage anything. I'm sure AMD's card will operate in spec, it's a new card with bugs, but apparently only AMD gets crucified for that, everyone else gets a pass.

AMD's card doesn't operate within spec. That's the point.
Had we not pointed this out, the company may not have ever addressed it.

To my knowledge, this is the first time we've discovered a graphics card that pushes the limits of PCIe so far.
It likely won't fry components, but specs exist for a reason and should be adhered to. The public backlash is wholly waranted. No one else is "getting a pass."

 

[salgado18]

Don't forget to test what exactly is this "Compatibility mode" and it's impacts. Please give us the whole picture, even if you publish it a bit later than other sites! It's public service, not a race 😉

 

[elbert]

Cool but is it true that some of the RX480's 4GB really have 8GB's that users can update the bios to access?

 

[wizout]

So far I've seen only 1 reported case of motherboard dying with a RX480 in it and there is no evidence the graphics card is responsible. It might very well be a regular failure.
Can anybody point me to the 'destroyed' motherboard people keep talking about?

 

[Sam Bittermann]

I'm sure everyone will test the hech out of the new driver and AMD haters will still hate AMD no matter what the results are. It's more fun to continue the band wagon of kicking a dead horse than one that's alive and can kick back.

 

[turkey3_scratch]

Can't wait to see the retest of this card once the drivers are updated! I also ponder what performance hits it will take, since AMD's response said it will only have a "minimal performance impact".

 

[therealduckofdeath]

If a motherboard has such a big design flaw that it can supply more current that its hardware was designed for, that would really be a motherboard issue. Yes it's a bad release by AMD, not thoroughly checking if it properly complies to industry standards but, IF a motherboard is fried because of this I'd actually complain to the manufacturer of it.

 

[turkey3_scratch]

If a motherboard has such a big design flaw that it can supply more current that its hardware was designed for, that would really be a motherboard issue. Yes it's a bad release by AMD, not thoroughly checking if it properly complies to industry standards but, IF a motherboard is fried because of this is actually complain to the manufacturer if that one.

A power supply can easily fry hardware with excessive ripple. Would it be the hard drive's fault if that causes the hard drive to fry?

 

[therealduckofdeath]

That is a nonsensical comparison, kcarbotte. The gpu does not supply or generate the power.

 

[therealduckofdeath]

(Edit: since edit doesn't seem to work in the broken comments forum, here's an old-school edit)

That is a nonsensical comparison, turkey3_scratch. The gpu does not supply or generate the power.

(Edit: To respond to the correct person 😀 Darn you mobile UI \o/)

 

[lucas_7_94]

Will be a new benchmark review with the new driver update? I want to see how much that 3% do.

 

[George Phillips]

I have been bashing AMD on why they didn't make the Polaris chips at TSMC who can probably make them run at 2GHz+. AMD should have quickly decided to abandon Global Foundry and go with TSMC next round. It might be saving money to let GF make the chips, but the end products are slower, hotter, cheaper, and less profitable!!

I really want to see Radeon become so competitive with Nvidia again in the high end cards.

 

[RazberyBandit]

Based on how the reference model is wired/traced, I don't think this can be fixed in software, with the possible exception being an updated vBIOS that forces the IR3567B to call for more power from the 6-pin PCIE connector. Sure, the so-called limit for a 6-pin connector is 75W / 6.25A, but most are fully capable of delivering three-to-four-times as much current.

It's the five tiny 12V pins within the PCIE slot that aren't built to handle more than 1.1A @ 12V that are truly at risk of burning out. These pins are what are at risk, not a potential (so-called) overdraw on the 6-pin PCIE connection.

 

[juggernautxtr]

Based on how the reference model is wired/traced, I don't think this can be fixed in software, with the possible exception being an updated vBIOS that forces the IR3567B to call for more power from the 6-pin PCIE connector. Sure, the so-called limit for a 6-pin connector is 75W / 6.25A, but most are fully capable of delivering three-to-four-times as much current.

It's the five tiny 12V pins within the PCIE slot that aren't built to handle more than 1.1A @ 12V that are truly at risk of burning out. These pins are what are at risk, not a potential (so-called) overdraw on the 6-pin PCIE connection.

Its called wattman..........(watt manager)
quite easily accessible through a driver tweak now. what you see is the basic side of the power management.
the base programming the things you can't see are most likely guidable through software tweaks.

 

[cmi86]

From what I understand the "going up in flames" comments are basically impossible. Burning out a PCI-e slot while extremely unlikely has been reported to have happened a couple times on old boards. Truth be told the only honor I remember for a PC component literally going up in flames belongs to Nvidia with the 480's ahh yes, the old PC incendiary grenade...