Oh? Seem like the displayport blackscreen issue.
Quite famous with 290X and 290.
Recent vbios somehow fixed it.

But apparently, some AIBs dont get the memo about latest vbios, so AIBs still flashes the same initial old problematic vbios for the 390x and 390 (rebrand is a rebrand no matter amd tries to deny it).

EDIT: Sapphire and ASUS are confirmed to use the latest VBIOS for the 'rebranded' 390x and 390.



This post has been edited by JohnLai: Jun 30 2015, 01:32 AM

Which AIB brand of 390x/290x and 390/290 that have problem in your case?

Edit: In case if you guys want to know, the beta VBIOS update from Sapphire disables the VRAM powertune functionality(only two vram power states) as well increasing a little bit voltage to GPU core and VRAM. GPU powertune is retained.

In ASUS vbios though.......asus decided to disable powertune functionality for both gpu core and VRAM in addition to slight increase in voltage for gpu + vram.

AMD actually knows about this issue, which is why the driver automatically load the "fix" once it is installed. But how the hell people install the driver if they can't get pass the black screen in first place. Not many know how to enable integrated GPU in order to install the discrete gpu driver. Those with AMD FX processors are screwed because the fx series doesnt have IGPU.

This post has been edited by JohnLai: Jun 30 2015, 11:18 AM

Unclean power from PSU? The VRMs at the GPU ought to filter its out........

In any case, this is AMD fault in the first place. AMD doesnt even officially 'recognize' the issue for the 290x and 290 is related to voltage issue.
AMD did acknowledge the issue by claiming it is fixable by driver and yes, it did.......

The core problem is the VRAM controller and GPU core voltages stored at vbios.

Yet, AMD still insist on saying 390x/390 isnt a rebrand while mentioning the stuff on new power management µarchitecture.
At least for 390x/390 series, amd is trying to push the updated new power management µarchitecture vbios to AIBs.
As I said, too bad some AIBs 'don't get the memo' about it.

Time to flash the latest sapphire vbios into the card through DOS (before the windows boot? ATIFlash USB?)

EDIT: Some card got dual bios, have you tried it yet?

This post has been edited by JohnLai: Jun 30 2015, 05:16 PM

Hmm....how about forcing the PCIE controller mode into 3.0 instead of auto? (adjust at bios?)
Some said this solved the problem?

For the beta vbios, you need to specifically request it from sapphire technical support site.

Huh? You can set RGB Full Range 0-255 at Nvidia Control Panel for quite a while now since last year.



In any case, guess AMD really had 30k unit of fury x available for year 2015 even though AMD denied this rumour.

But there are many report around the net that customers are not able to purchase fury x.
Some even got their order delayed until July and October. Few of them got their fury x order delayed to December.

Air cooled edition will be released on 14 July,so I can only guess AMD plan to spread out the 30K unit evenly for June to December.

Since Fury Nano is cutdown based on Fiji core as well........uh oh, feels like a foreboding.

Very unlikely, the shortage is real.
June to December 2015 is the best time for AMD to start raking profit from AMD Fury X series and yet, AMD can't even provide a steady stream of product supply in the market. (Fury X was delayed several times before it was launched).

AMD said the company shouldn't be viewed as value brand, I beg to differ, I believe the only way for AMD to survive is to become a value brand.
If the company doesn't have the capability to produce high end product, then the correct course of action is to produce mid end or low end product.
This is exactly what happens to AMD right now. AMD aims for enthusiast range gpu, yes, there is a great demand in this market segment, but AMD fabrication partner DON'T have the capability to produce sufficient quantity of the said product.

Not to mention the barrier of entry into enthusiast segment is quite high.
AMD competitor, Nvidia, have a great deal with game publishers to use its proprietary gameworks and physx. These middleware actually starts to gain traction in the developer market (sad, but it is a fact). Meanwhile, AMD proposed physics engine never got any traction to move on.
AMD proposed Mantle doesnt do much right now when discrete gpu market is currently dominated by nvidia.
AMD Trueaudio chip doesnt gain much game developer interest as well.
The last knucker barrier of entry into enthusiast market segment is that Nvidia is pouring large amount of R&D money into their chip development.
In most cases, the R&D done by nvidia can be easily recouped back. The same can't be said for AMD Fiji. HBM sounds good and perform well, but can AMD recoups the initial investment? Or nvidia will reap the benefits of AMD hbm instead?

Fury X was originally slated to debut with price range of USD800 until nvidia decided to crash the party with Titan X cut down, 980Ti, setting a performance target at price of USD 649.

Why is Nvidia proprietary stuff gaining traction? Simple, because Nvidia is actively encouraging developer to use it by providing incentive (in term of finance and branding).
AMD tries to follow the similar method.
However, AMD financial situation is not doing well compared to Nvidia in addition to AMD brand doesnt carry much value in its branding. (Sigh, I am saddening by this, but the fact stands)

Doesn't look like AMD has sound strategy to get into enthusiast market segment right now.

Not me lah, I just discussing the Fury shortage issue. (I outta budget at the moment, need to save up to pay for ACCA membership fee and examination fee soon, RM drop like hell compared to Britain pound )
I saw many reddit + overclockers.net thread commenting on this issue and one from Phoronix

LOL , I saw RapeLay icon.

RM exchange rate drop too much, plus all electronics good is now subject to GST.

In year 2013, the exchange rate is RM2.9 to USD1.

Now, it is RM 3.8 to USD1. Then need to pay GST 6% on top of it.

Where is the contest?

TO AMAZE ?
Err....what did I miss?

Thanks

A common problem with factory overclocked ASUS r9 280x.
Especially if your mode is http://www.asus.com/Graphics-Cards/R9280XD...specifications/

Asus uses GDDR5 rated for 1500mhz for the board, but overclocked it to 1600mhz.
To make it worst, https://www.techpowerup.com/reviews/ASUS/R9...u_II_TOP/4.html shows the VRAM chips arent actively cooled.

You only need to downclock the vram clock to solve it. No need to downclock gpu core clock.

It is VRAM issue.

GPU Tweak? Tuning? Memory Clocks ? Set the slider to 6000Mhz? Apply? Save profile?

If I am not mistaken, you better update the VBIOS as well, I read at rog forum the VBIOS either pump more voltage to keep the vram stable or reduce the vram clock.
015.042 vbios version?

EDIT: Link to rog thread here!, read entire 64 pages if you want to know more

Let just hope your vram isn't damaged by the so called 'factory tested overclock' yet....

This post has been edited by JohnLai: Jul 16 2015, 03:43 PM

????
The bolded hertz is referring to VRAM clock or GPU core clock?

Official specification from Asus is ;
GPU Boost Clock = 1070 MHz
GPU Base Clock = 970 MHz

EDIT: I suggest you to update the vbios and report back if it is still happening.

This post has been edited by JohnLai: Jul 17 2015, 10:19 AM

Hmm....if it is VRAM clock, wouldn't setting back to 1500Mhz (Hynix chip original frequency rating) will solve it for his gpu?

He ought to update his vbios first before playing around with vram and gpu clocks.

As Skidd said~

Now, faster update to the latest vbios and report back if the issue still appearing.

=.=...........
Dont worry about bricking modern GPU vbios. You can always reflash back.
He have integrated gpu, no?

If vbios update failed, just turn on the igpu. Plug the display cable to igpu port and flash back the 280x vbios.

You really wanna know?

Apparently someone at rog forum measured the 280x default (before vbios update, not sure with the new vbios yet) voltage of 1.6V for VRAM voltage.

But, the asus 280x hynix chip is rated for 1.5V.... https://www.skhynix.com/products/graphics/v...lNo=H5GQ2H24AFR

Specification also stated the VRAM maximum temperature is 70 Celcius....so..........

1.6V + 1600Mhz(quad pump 6400Mhz) + not actively cooled (only fan blowing air?) = your current result.

Overheating is nothing strange......

This post has been edited by JohnLai: Jul 17 2015, 11:59 AM