AMD Releases Statement On Radeon RX 480 Power Consumption; More Details Tuesday
by Ryan Smith on July 2, 2016 5:05 AM EST
Following up on this week's Radeon RX 480 launch, there has been some questions raised about the power consumption of the card. This is after some sites whom directly tap the power rails feeding the card discovered that at least some of their samples were pulling more than the standard-allowed 75W over the PCIe slot and/or 6-pin PCIe external power connector.
To that end, it would appear that AMD's staff is working weekend duty, and they have just sent over the following statement.
As you know, we continuously tune our GPUs in order to maximize their performance within their given power envelopes and the speed of the memory interface, which in this case is an unprecedented 8Gbps for GDDR5. Recently, we identified select scenarios where the tuning of some RX 480 boards was not optimal. Fortunately, we can adjust the GPU's tuning via software in order to resolve this issue. We are already testing a driver that implements a fix, and we will provide an update to the community on our progress on Tuesday (July 5, 2016).
If some of the data is to be believed, these cards are exceeding 150W total at times, which would mean there is either something causing them to run in the wrong power state, or they are just outright exeeding their power limit and need to be throttled back. As we don't do per-rail testing I don't have anything meaningful to add at this second, but it will be very interesting to see how AMD responds next week.
Update 07/06: AMD has since released their status update, which you can find here.
Facebook
Twitter
RSS
181 Comments
View All Comments
D. Lister - Saturday, July 2, 2016 - link
"My assumption is that fixing a problem with transient spikes is hardly in the same league as the 970 VRAMgate."Your statement is correct, although the problem here isn't transient spikes, but a sustained overload. Still, unlike that 970 nonsense by Nvidia, this is fixable, at a slight performance cost.
Yojimbo - Sunday, July 3, 2016 - link
I think you guys have things backwards.The VRAM situation of the 970 had no effect on the performance of the card. The benchmark results published by review sites were all accurate and true. After the issue came to light, despite much searching, no one could verify instances under playable frame rates where the card was hamstrung by its memory configuration. Why? Because the memory system of the card, as it was, was engineered to be that way, and the card was well-engineered and balanced. It wasn't memory bandwidth or memory capacity bound under playable conditions. The only error was not properly informing the review sites of the memory architecture.
This RX 480 power draw controversy is a completely different, and much more serious, situation. AMD also failed to notify people of the situation, a situation which includes the total power draw of the card exceeding the TDP for extended periods of time and the power draw through the PCI-E slot exceeding the specifications of the standard (whether they, AMD, knew of the situation or not). However, the situation is both potentially damaging to a user's system and potentially invalidates the published benchmarks of the card. We'll know more in the next few days how much of an issue the latter is, but we already know the former is a concern.
D. Lister - Sunday, July 3, 2016 - link
"The VRAM situation of the 970 had no effect on the performance of the card."My point was merely that this issue, problematic or not, was not something that could be sorted out with a patch, without redesigning the chip, or at least its memory subsystem. Which I suppose they eventually did, with Pascal.
"This RX 480 power draw controversy is a completely different, and much more serious, situation. "
No argument there (please see my earlier post http://www.anandtech.com/comments/10465/amd-releas... ). The problem is serious, but it is most likely fixable with a software update.
Yojimbo - Sunday, July 3, 2016 - link
But the 970 didn't have a problem at all. The only mistake made was a misrepresentation of the specs handed out in the press packet. If the correct information had been handed out people would have said "oh, that's clever" and hardly anyone would have given a shit.vladx - Saturday, July 2, 2016 - link
If there was a similar problem with nVidia cards it would've been reported by now, just like the '3.5GB' debacle was discovered and shown to the entire world.Oxford Guy - Sunday, July 3, 2016 - link
As I recall, it took several months for the enthusiast community to discover the problem. Meanwhile, Nvidia raked in cash from "4 GB" 970 sales. It seemed to be a particularly good value for SLI at the time since it was supposed to have the same amount of high speed VRAM as the overpriced 980.After enthusiasts discovered the performance inconsistency on their own, Nvidia admitted the truth via this site -- which ran an article (where it was claimed that the extremely difficult to believe story about the engineering team coming up with the design all by themselves and not telling anyone about it is credible) but still refuses to list the proper specification on its website.
brunosp - Saturday, July 2, 2016 - link
No problem https://www.youtube.com/watch?v=elgIUK0ond0brunosp - Saturday, July 2, 2016 - link
http://wccftech.com/article/radeon-rx-480-reducing...tipoo - Saturday, July 2, 2016 - link
Interesting, I hadn't known it only impacts the 8GB and therefore 8GBps cards before. Did people test the 4GB and find there was no issue? That's probably why there was so much confusion over whether this was regular or not.And any changes to performance after the patch will definitely be interesting. Even if it goes fine though, what a PR mess.
HomeworldFound - Saturday, July 2, 2016 - link
It's not that much of a PR mess, only as big as people like yourself make it. In six months time it'll only matter to people trying to justify their purchases to themselves and to those people that spend more time arguing over which company is better than they use their graphics card.