• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

The AMD Driver Thread

UTmaniac said:
Just had the black screen with white (active) BenQ power button issue again, then a reboot ~30 secs or so later, after clicking to change Chrome 64-bit tabs.

I bought a better DP cable from Amazon at the start of the year, after reading in here that the supplied DP is ancient spec.

Recently when this issue started, I even tried swapping DP ports on the back of my 390, no change.

Having rebooted, I've turned off hardware acceleration in Chrome 64-bit settings. Has this fixed things for others?

I have Sapphire Trixx running at Windows startup, to force a fan speed of 75% currently. Could this be conflicting with AMD drivers and causing the black screen issue above?
Click to expand...

What drivers are you on? I upgraded to 16.8.3 yesterday and haven't had a black screen since then.
 
hardjack79 said:
So I've now tried a different cable between different DP ports at both ends and just had another blank screen. I don't think its the cable... other than perhaps a more fundamental problem with the DP link between GPU and monitor. Have lowered refresh to 60 Hz now, will see. Still confused why this happens only with the new monitors and not old -- could be a DP thing, a bandwidth thing (or res or refresh individually), a voodoo curse..

I do have modular PSU, but as this isn't happening under load I don't think that would be the issue. ****, you never know.

I'll put the 480 in the other PC if I can't isolate the cause of this soon. I need to know whether to RMA the screen or the RX480 soon because I move to NZ in a month. :eek:
Click to expand...

I tried running the screen on DP at 60 Hz and didn't get any black screens. So I'm tempted to say that this is partially a bandwidth/load issue. And because it only happens during the lowest power state on the GPU, I suspect that it just can't handle the additional (4x) load that is required compared to 1080p60. I tried to boost the voltages a fraction, but the lowest power state is the only one that can't be edited. :(

Anyway, after that I upgraded from 16.8.2 to 16.8.3 to see if that would let me change the voltages and, no it doesn't, but I haven't had a crash since then (10 hours running time). In the change notes AMD do say:

Fixed Issues
A random blank or colored screen may be experienced intermittently on some Radeon RX 400 series products while gaming.

Other than the "while gaming" bit (which I suspect is just a force of habit for AMD change note writers) this was the problem I was having.

If anyone else is having the same issue, then 16.8.3 might be the fix.
 
Is there a way to fix the core speed to 1050 on a fury nitro while gaming? As its all over the place and I'm noticing it when gaming with FPS are all over the place, i've set power limit to +50 and its not going over high 60's as I've set a fan profile on afterburner. I've tried a few games and they contantly drop under 60fps (freesync enabled), sometimes around the 30fps. I've googled it and it appears to be quite a common issue, but haven't seen a fix yet
 
Although it looks like the main fix is to use Clockblocker, is this recommended? Seems a bit odd that you have to install a non amd program to get it to work properly
 
When they added the Power Efficieny toggle into the drivers, that fixed it for me, as i used to suffer from the constant core clock bouncing.

Id recommend Clockblocker if that or Afterburner doesn't work, as plenty use/used it to fix it.
 
So whats the deal with these load spikes?

Wh2wz1g.jpg.png

This is playing BF1 on Ultra settings above (load gap with nothing is me tabbing out)

Should be solid right? Happens in other games too.

This 480 is becoming a real pain in the arse.. fast.

This 480 even has Task Manager wanting to use it, when idle, if i bring up Task Manager i see my core clocks randomly spike up for a split second, i even see fan speed spike to 100% sometimes though i dont hear it ramp up.

Is there a previous driver without all these issues me and others are having? Especially the 100% load bug.

I'll roll back a driver or 2 if needed, i only had this card since 16.8.2 which i didnt notice any problems with (doesnt mean they wernt there) until i installed 16.8.3.
 
Are you actually seeing any performance issues? AMD GPU load has always looked like that in GPU-Z and Afterburner for me, at least from Hawaii onwards. The Fury that I had recently was exactly the same. Spikes from 0-100% and back again all the time, but actual performance was fine. I always assumed it's just a case of them not reading the load correctly.
 
Aretak said:
Are you actually seeing any performance issues? AMD GPU load has always looked like that in GPU-Z and Afterburner for me, at least from Hawaii onwards. The Fury that I had recently was exactly the same. Spikes from 0-100% and back again all the time, but actual performance was fine. I always assumed it's just a case of them not reading the load correctly.
Click to expand...

Looks like that on my cards too. Havn't noticed any issues.

Suspect this really needs AMDmatt for a definitive answer though.
 
Last edited:
Set my 480 to primary monitor only and all these issues go away, running my other 2 from my 6700K iGPU now.. as long as i play video etc on the iGPU monitors its all good.. core clock is settled now even playing video on primary though memory clock ramps sometimes with temps up a bit, before the core clock would be hitting every number above 300.

Suppose i'll leave it like this until AMD fix their crap.

At least Intel know what they're doing when it comes to iGPU.
 
N19h7m4r3 said:
I'm on FreeSync with the Asus MG278Q and Fury X, none of the mentioned issues either. Not even the Freesync flicker.

*Touch Wood* None of it has affected me and I'm really happy about it.

Even so I hope AMD sorts it out rather quickly, they really cannot afford to have issues like this.
I can only image now badly that flicker can affect someone with Photosensitive epilepsy.
Click to expand...

That's good to know, I think I'm going to go for a fressync ultrawide soon seeing as the wait for Vega is expected to be well in to next year.
hardjack79 said:
I'm a kiwi but haven't lived there for 15 years.
Click to expand...

It's a beautiful place, I've never been but that much is obvious, My nephews coming back from 2 years in Oz at the end of the month but I don't think he's that happy to be coming home, I wouldn't be :)
 
hardjack79 said:
I tried running the screen on DP at 60 Hz and didn't get any black screens. So I'm tempted to say that this is partially a bandwidth/load issue. And because it only happens during the lowest power state on the GPU, I suspect that it just can't handle the additional (4x) load that is required compared to 1080p60. I tried to boost the voltages a fraction, but the lowest power state is the only one that can't be edited. :(

Anyway, after that I upgraded from 16.8.2 to 16.8.3 to see if that would let me change the voltages and, no it doesn't, but I haven't had a crash since then (10 hours running time). In the change notes AMD do say:

Fixed Issues
A random blank or colored screen may be experienced intermittently on some Radeon RX 400 series products while gaming.

Other than the "while gaming" bit (which I suspect is just a force of habit for AMD change note writers) this was the problem I was having.

If anyone else is having the same issue, then 16.8.3 might be the fix.
Click to expand...

Amd have had this problem since gcn 1.1, it ****es me off, it's to do with 2d clocks p-state switching, and effectively how the memory voltage is applied when the 2d load changes. It's why people think turning off hardware acceleration is the cure, when it's not it just means the cpu is decoding the video. 60hz or single screen 60hz doesn't require the memory clocks to bump up, but dual screen or 144hz will.

Does us a favour do some screen prints of gpu-z of your screen/s and when set to 60hz and 144hz. I'd like to see if the memory clocks raise to full speed when over 60hz.
 
36
down vote
+100
ATI drivers had mouse corruption errors since 2001.
You have already discovered almost all the existing workarounds.

You should contact their Support, and in the meantime keep on looking for driver updates. But I don't think that a 12-year problem has any hope of an early solution.

Below is the complete list of workarounds that I was able to find :

Enabling/disabling shadows for mouse
Enabling mouse trails
Disable Catalyst AI (through ATI Tray Tools)
Disable mouse shadow
Disable Windows Aero
Use a third-party cursor app (cursorFX, etc.) to draw a better cursor
Return the DPI setting in Windows 7 to normal (if changed)
Change the dual monitors positioning in Windows (use Above etc.)
Do not use ATI for multiple monitors
Move mouse to lower right corner, completely
Move mouse right under a taskbar item (lower most), and shake it left and right
And here's a possible list of ways to "fix" the cursor without rebooting:

Unplugging mouse
Open magnifier and zoom to 100%, leaving the magnifier open
Activating an eyefinity profile (for those with multiple screens)
Changing resolution
Put Windows into sleep mode
As soon as you see the stripes/comb pointer, move the mouse to the edge of your screen (top right was mentioned) and shake the mouse a few times

Since 2001?! damn
 
Is it something that triggers this glitch? Since owning Amd I never had this happen to me? I only use single display, used 60hz, 120hz and 144hz. Games on a lot of different display resolution over the years. Even crossfire.

Too the guys that have had this what is your system specs?
 
You must log in or register to reply here.
Back
Top Bottom