That could cause a huge jump in power consumption. Either GPUz is wrong (needs an update for Polaris?) or there is a bug someplace.
http://www.pcinvasion.com/amd-looking-rx480-pcie-compliance-failure-reports
A user on reddit also posted the GPUz shot of his RX 480 under gaming load and it was running at 1.29vcore, well above the 1.15v max allowed voltages under Wattman.
This looks like a Crimson problem failing to regulate the power and running some cards over-volted.
Yeah, I'd love to read an article about that.No, it quite a lot more involved than that. There are very detailed test procedures that must be adhered to. Nothing about standards compliance is simple.
I've read the PCI spec several times and still haven't found the 75W number. I don't really care either way, but how did that number come about again?
No, it quite a lot more involved than that. There are very detailed test procedures that must be adhered to. Nothing about standards compliance is simple.
You have access to the PCI-E spec?
Well, it makes no sense to send any cards to reviewers that might have this problem.
One would think that AMD would check the cards before giving them out to make sure there are no such surprises.
So, I have to conclude that somehow AMD was unaware of the possibility of some cards/chips being out of spec.
That could cause a huge jump in power consumption. Either GPUz is wrong (needs an update for Polaris?) or there is a bug someplace.
The cards reviewers got all had a 4gb/8gb spec where reviewers could flash the cards and test for either/or configuration, so AMD wouldn't have to send 2 cards each.
These aren't the same cards sold to consumers (physically limited to either 4 or 8gb, I believe). Is it possible that there is something janky in those boards, concerning the "flashing" or whatever is needed to switch from 4 to 8gb that is causing unaccounted-for power draw issues?
I don't really understand any of this, but afaik, none of the consumer cards have actually been tested.
Huh, I guess the specs are easier to find than I thought. You need to be a member to get the docs from PCI-SIG's website.
Anyways, you need to look at the PCI-E Card Electromechanical Spec. This looks to be a dated REV, but I don't think power specifics have changed since.
Page 36, 4.2 Power Consumption has info on the 75W max slot requirement. Including note 4, which states "A x16 graphics card is limited to 75 W. The 75 W maximum can be drawn via the combination of +12V and +3.3V rails, but each rail draw is limited as defined in Table 4-1, and the sum of the draw on the two rails cannot exceed 75 W."
Well then we should probably not rely on these reviews at all.
I thought it was bad to send reviewers unique cards?
Except that the power variation across several reviewers already makes your hypothesis impossible. If AMD loaded special BIOS they would all be more or less consistent. Link to the Reddit megathread from the dude pushing the fails PCI-E spec. Also confirms that some reviewers used retail samples for testing.
https://www.reddit.com/r/Amd/comments/4qfwd4/rx480_fails_pcie_specification/
This could only be an issue for miners. Gamers are safe
Yeah but that's super old (think you would agree) well at least in the tech industry. The updated guide basically states you can have over 75w if I'm reading it right. Basically 25w, 24w, 23w, 22w. As long as the less than the greater. So a total of 94watts (just an example).
6.16 is where I'm reading.
You need to be reading PCI Express Card Electromechanical Specification Revision 3.0, if you can find it. I can't, but that will tell you the power consumption for slot and connectors.
The cards reviewers got all had a 4gb/8gb spec where reviewers could flash the cards and test for either/or configuration, so AMD wouldn't have to send 2 cards each.
These aren't the same cards sold to consumers (physically limited to either 4 or 8gb, I believe). Is it possible that there is something janky in those boards, concerning the "flashing" or whatever is needed to switch from 4 to 8gb that is causing unaccounted-for power draw issues?
I don't really understand any of this, but afaik, none of the consumer cards have actually been tested.
Yes, it confirms what I've been thinking all along.
A PCI Express 150W-ATX add-in card is defined as a card that consumes greater than 75 W with support for up to 150 W inclusive. A card that uses a single expansion slot is described as a SINGLE-SLOT add-in card. A card that extends into the adjacent expansion slot is described as a DUAL-SLOT add-in card. A card that extends into the two adjacent expansion slots is described as a TRIPLE-SLOT add-in card. A 150 W add-in card, as with any CEM add-in card, may be SINGLE-SLOT, DUAL-SLOT, or TRIPLE-SLOT. A system that supports a PCI Express 150WATX add-in card is required to ensure that sufficient power and thermal support exists. For example, in an ATX form factor system, the adjacent expansion slot can be left vacant allowing for 1.37 inches of clearance for the add-in card as illustrated in Figure 1-3 to support a 150 W or lower power DUAL-SLOT add-in card.
A PCI Express 225 W add-in card is defined as a card that exceeds PCI Express 150 W 1.0 power delivery or thermal capability and, as such, consumes greater than 150 W with support for up to 225 W inclusive. This card, as with any CEM add-in card, may be a SINGLE-SLOT, DUAL-SLOT, or TRIPLE-SLOT add-in card
As long as the designed cooler takes up two physical slots, it's allowed to use more than 75W.
Yes, it confirms what I've been thinking all along.
As long as the designed cooler takes up two physical slots, it's allowed to use more than 75W.
From what I've read, some reviews, like TPU, got basically retail version of cards but with a BIOS for 4GB/8GBs and tools to flash it. The Stilt commented on this in the CPU section.
Chances are cards in the wild are going to experience some randomness. I hope for AMD's sake, more people get good cards than lemons.
Makes me wonder if they are binning all the good chips for AIB.
Yeah, i am really wondering if AIB card reviews are suddenly going to look much better in power usage simply because they will be retail cards.