Discussion: Polaris, AMD's 4th Gen GCN Architecture

Page 36 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Status
Not open for further replies.
at this point i am just sitting back and waiting to see what happens. no need to get the panties all in a bunch when it is out of our hands. just be aware of the possible issue and see how it ends. if you have the card, i'd suggest not oc'ing it for now until the issue can be addressed/fixed/disproved or whatever else is gonna happen. i see some other sites looking a bit further and seeing the same issue and others not seeing it.

clearly it is not a problem for 100% of the cards. maybe they can get lucky and identify a particular production run and take some pro-active steps. so far i have not seen any info on what the specific cards are with the issue. are they all amd made cards or are some of the other brands the culprit or is it an issue with the design that can pop up on any of the boards made to this design?? so much we don't know and jumping to conclusions one way or the other is not gonna do anything for any of us.

as Aaron Rogers said "RELAX" it's not over yet.
 


Not like we can do anything anyway 😛
 
Maybe a fix is on the way? (Much more info in the link, if you haven't checked it out already.)
For its part, AMD has acknowledged the issue publicly during its launch day AMA on Reddit and to me through email. They have stated they are looking into the issue and may even be able to come up with a software-based fix. I hope that's the case and I'll be curious if there are any performance changes that occur because of it. Consider this space reserved after a more thorough discussion with the Radeon team.
http://www.pcper.com/reviews/Graphics-Cards/Power-Consumption-Concerns-Radeon-RX-480/Overclocking-Current-Testing
https://www.reddit.com/r/pcmasterrace/comments/4qfy9d/i_work_at_amd_the_time_has_come_to_ama_about/d4smb2m
Click to expand...
 


When I set Voltage control to manual on my XFX RX 480 4gb, my voltages are already set the same or lower than what they say in these articles.

Maybe thats why mine hasn't blown up yet, or different cards have different default settings
 
I'm wondering if this is just a manufacturing defect as opposed to a mass problem.

AMD responding or not is a no win. If they respond now and say "we're investigating the reported issue" then people will blast them for not moving fast enough. By not responding, people think they are ignoring the issue and don't care, blasting them for it. Either way ends badly for AMD. I like that they haven't said much until they have something to informative to say. However, the responses the marketing guy did on redit was a mistake and makes it look like they are tone deaf...
 


Maybe it only affects 8GB variants?
 


This case is a bit different. When you talk about overall consumption, they will overlook it because it seems like it's a recommendation, but the PCIe 75W cap is a requirement. Or I understand it like that.

In any case, it remains to be seen what AMD finds. Don't recommend the card until we know more; I agree with that.

Cheers!

EDIT: http://www.anandtech.com/show/10465/amd-releases-statement-on-radeon-rx-480-power-consumption

Ok then, on Tuesday we will see/know if it's fixed or not.
 
"...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..."

Sounds promising.

Hopefully this will not impact performance of the reference cards in any significant way. I'm waiting for the custom cards anyway - as I always do.
 
I have to say I noticed my computer at idle is bouncing between 150 and 200 watts. It NEVER did that before. Also if I look in that new Radeon Wattman thing I see that occasionally the clock and memory clock jump up, on the desktop. Same thing never seen that before. My old cards just sat there at 300mhz when not playing 3d all the time.

I would hope a driver can fix this, its very strange.
 
Idle clocks is definitely the kind of thing you'd expect a driver update could fix. Both AMD and Nvidia have had instances in the past of certain stuff making their GPUs needlessly jump to higher clocks. Changing the basic power consumption at load is more tricky, short of making the card throttle.
 


I was watching a video on OC3D's Youtube channel regarding overclocking the RX 480, very odd card. If you want to overclock, all you do is up the power target to max, don't change the Mhz and you'll get far better numbers. But then you get into thermal throttling with the reference cooler. At least EK is shipping waterblocks out...

I think the RX 480 still has hope in aftermarket coolers. Hopefully they will not only A, put dual 6 pins or 8 + 6 pins on the cards but also B, up the power target by default, it's going to be a really good monster. Probably equal to 1060 performance.

But, I'm defiantly not buying one for my personal rig. I'm getting the gtx 1060.
 
I am curious how well the reference 1060 cooler will be since it's cooling won't be as good as the higher end FE cards. There is no vapor chamber.

It's gona be a while before I buy one though, I want the 6GB model (why would nvidia launch a 3GB model and this card is going to be sorta designed for VR?). But $300 is a lot of money, I hope prices fall in September when I can actually afford one.

GRRR, Techy you dummy. Sorry folks, now I'm derailing this thread. So please don't reply to this. Or if you do, make sure you bring it back to AMD somehow.
 
seeing amd's response, i sure hope a driver fix is all it needs. or even a bios update would be a simple enough thing for most users to accomplish.

seen a bit about undervolting the card and it seems like it may be similar to the 300 series. if you recall, undervolting them a bit actually helped performance a lot. temps and power draw was way down and performance went up a notch. seems likely amd did something similar this time around as well to ensure more chips were usable. wonder if tom's can do a follow up for us with the card and test undervolting once this new "software fix" is in place as well as the obvious power draw and oc testing that needs to be redone.
 
something tells me they will do a new test. just seems the responsible thing to do for the readers 😀 hopefully as quickly as possible for those who have already bought the cards. as i said above, i'd like some undervolting to be thrown in this time around as well. might prove to be worth doing and i like how well they test with that nice high end power equipment.
 
Status
Not open for further replies.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom