INFORMATION REQUIRED WHEN POSTING A QUESTION Full system info,please.
1.Delete C:\AMD folder
This does not work. After doing this the black screen error happens as soon as the password screen pops-up to login to Windows 10. The GPU seems to be loading faster and therefore crashing faster.
Thank you redfury, I will try that soon
Describe your system:
- AMD Graphics Card
- R9 390
- Desktop or Laptop System
- Operating System
- Windows 10 64bit
- Driver version installed
- Radeon Software Crimson Edition. Latest edition, (ill have to check when I install again).
- Display Devices
- AOC 2436 Monitor. HDMI cable
- MSI 970A-G46 (MS-7693)
- AMD FX-8150
- Power Supply Unit Make, Model & Wattage
- CoolerMaster SilentPro 850w
- 8GB Corsair Vengeance
I've updated my BIOS within the last year actually so I have a good BIOS. I just checked to make sure and all my BIOS is up to date.
You may want to get ahold of Msi on this one just found this with he link t your bios
* Please download and execute the file in USB pen drive only. System will restart several times to update system BIOS with NVRAM & ROMHOLE. Please DO NOT remove the USB pen drive until BIOS updated successfully. * Please make sure you have read the SOP before you update the BIOS. * MSI recommend you to update BIOS only if you are using Windows8. * There is no way to reflash back to BIOS version 1xx if you already updated to version 2xx.
The date on it is //2016-01-08//
- AMD Graphics Card
Do you get this crash without installing any VGA drivers, say after you installed windows-10 or / uninstalled the VGA drivers and sit at desktop, and wait to go idle.
Does the problem not happen ? or do you instatly get the black screen again and your thinking its the GPU or the new Crimson drivers.
What i picked up on is you moved from windows-7 to upgrade to windows-10 . Did you buy the mobo for windows10 or was this your exact setup when you were using windows7.
AMD 970 chipset based
Problems i see with this your board SLI and Crosfire, something windows-10 has decided not to support fully and Nvidia already pulled te SLI adapters on newer cards.
APS (Active Phase Switching) is win10 still supporting this, and did you find real 1607(win10 drivers for this bard feature ?
OC Genie II – Overclock & Unlock CPU Core in 1 Second ) Di you overclock in windows7 at anytime, and make sure to flush/ reset all features and board back to factory non-OC defaults, have you installed any of the OC features into windows10 , your getting alot o stuff like WattMan , Relive tons of features installing Crimson ...
Are they supported on your board or is that a conflict ?
Just trying to give ideas and help
I have a ton more if interested.
No problems in Safe Mode (with either GPU)
No problems with older GPU
(AMD Radeon 6900 running AMD Caralyst 15.7.1 for Win 10-64)
No overclock (that I am aware of)
no crossfire (only 1 GPU at a time)
With AMD Radeon 390 GPU, I can install hardware, get to password screen, boot in safe mode, uninstall drivers using program, reboot.
But after, the drivers are still self-installing (even though Display Driver Uninstaller said it would prevent this!)
My resolution changes to 1920x1080 and soon after my screen goes black...
The BIOS is up to date according to MSI Live Update. I had updated it since the last Jan 2016 update.
When I bought the AMD Radeon 390 GPU, I had to update the BIOS to the latest version for my GPU to work.
I updated to windows 10 about 8~ months ago.
I installed my AMD Radeon 390 GPU about 10~ months ago.
The Black Screen problem started for me 1-2 weeks ago.
The only changes to my computer recently are
- GPU driver updates (but reverting to last driver does not fix)
- Windows Updates
1.Restore your system to restore point (to a date earlier than 07.01.2017)
2.Try to boot in low resolution mode and see what will happens
3.Reinstall OS(i suggest you win8.1) or perform win10 FULL reset
This is a bit extreme. I don't have a windows disc so i cannot reinstall OS, and the issue is exclusively with the GPU not the OS anyway. It might be something with the compatibility between my BIOS and GPU.
Just a thought, try changing the Generic Driver named ( PCI Express Root Complex ) into ( PCI Bus ) and allow AMD to install onto the the PCI Bus into the AMD Pci-bus
If its reverse , then do the opposite and just try to get the new built for Win10 Pci bus to install over what ever is replacing it.
Im not sure which bus win10 installed for your system or how all your hardware picked up when it fresh installs, but if anything , a clean install of the OS with all hardware off would be a good idea, like i said im not how your system would come up with ALL OFF or how it comes up with all features ALL ON,
But this is defiantly sounding like Win10 installed a Generic driver somewhere it shouldn't have, because it just didn't know to list it unknown or picked something not-so-close to what it or they really are.
I know most people are counting on installing drivers and having them replace other drivers, but with Win10 if something say Inferred device is really something GPU related and it installs ontop of that, you would be looking at hardware that never really got configured correctly or got identified in the system improperly.
It is after all a really old board, another Thought (Non-volatile random-access memory (NVRAM changed with 1607 anniversary edition ) Manufacture came up with what was Win10 Ready for legacy / older system using tests based on public release. im still looking for the official announcement i read, but my bookmarks are a mess, takes a computer to use a computer, someone pass that on to microsoft maybe they will wake up (imho)
Hi, you can't do anything to avoid the black screen errors on the R9 390. I believe it's a driver issue and it won't be fixed ( i lost hope ). The best workaround = Downclock your GPU from 1500mhz to 950mhz, problem solved, no more black screen doing 2d stuff.
I looked into Radeon R9 390 more. Apparently there are a lot of people who get the black screen error
R9 390 and black screen problems
I did another test to install an old 'Catalyst' 15.11.1 driver instead of the 'Crimson' editions that seem to be causing the problems. I thought this would work but the black screen happened again after installing the drivers and rebooting. I think there may be some kind of auto-installing of crimson stuff behind the scenes that I don't know how to stop.
I'm going to try this last thing of down-clocking my GPU ... Though this seems totally insane thing to have to do. It's looking like the best bet is to replace the GPU and at this point I'm not very confident in AMD.
Using MSI Afterburner to limit to 950mhz kind of works, i still get black screen but it flickers back to the screen sometimes, but sometimes it still goes totally black too. Also I have to wait and pray that i get to the startup point where afterburner loads before the black screen error occurs. This is just not functional.
No problem so far with my 390.(new, about 3 weeks), have tested from 16.11.5 - 17.1.1.
Maybe you have a faulty Vbios, contact manufacturer's support.
I was fine too, for about 8 months. There are a ton of people posting about this issue, so its not an isolated case. Do you have a newer motherboard? Maybe your BIOS is better.
After 1 month, mine started blackscreening while watching youtube, twitch or just idling on desktop. Trust me, the only way to get ride of it is to downclock it, i ve been doing it for like 7 months now, 0 black screen.
Can you please explain the details of downclocking? Using MSI afterburner, downclocking to 950mhz seems to cause slow 'flickers' where it recovers from the black screen, but this is still annoying. Also there are several GPU settings, what are the exact specifications to set MSI afterburner to?
Sorry for not responding so quickly, i was kinda busy. Well i just set the memory clock to 950mhz, i do it in radeon settings. sometimes i get a blackscreen if i try to change it in msi ab. I have the sapphire r9 390 nitro with backplate.
"Well i just set the memory clock to 950mhz, i do it in radeon settings."
I checked all of my Radeon settings and i dont see any way to downclock my GPU memory. I'm currently using 16.11. Can you explain the steps to downclock in radeon settings?
What error codes/messages show in windows reliability history/monitor(not event viewer).
Hw acc. turned off. How many start up/minimized app/programs running.
A fresh install of win10 wouldn't hurt, tenforums.com/tutorials/9230-windows-10-iso-download.html
nothing of note usually in reliability history. For example the HDD went to black several times today, but there are no errors in the log. There are several notificaitons because i have been swapping cards and drivers several times a day. Here's on thing of note from 1/20
Faulting Application Path: C:\Program Files\AMD\CNext\CNext\RadeonSettings.exe
Problem Event Name: BEX64
Application Name: RadeonSettings.exe
Application Version: 10.1.1.1666
Application Timestamp: 5879075a
Fault Module Name: StackHash_ac46
Fault Module Version: 0.0.0.0
Fault Module Timestamp: 00000000
Exception Offset: PCH_1A_FROM_ntdll+0x00000000000A6C24
Exception Code: c0000005
Exception Data: 0000000000000008
OS Version: 10.0.143126.96.36.199.768.101
Locale ID: 1033
Additional Information 1: ac46
Additional Information 2: ac462b2fd07cc853b652b5104a1cf381
Additional Information 3: 189e
Additional Information 4: 189e0906de0aacba3aece588bfcc9060
Extra information about the problem
Bucket ID: 66062a413b0a277b3b0d6afe7bb0d54c (116318232397)
Very minimal programs on startup. Norton, Printer Manager, Audio Manager and the AMD control center.
I'll look into that Windows ISO that is good to have, but id prefer not to reinstall windows, and it doesn't seem like it has anything to do with windows anyway. My older GPU works totally fine and this new GPU was working fine for months on Win 10.
not sure what you mean but i have tried 4 different drivers including the old Catalyst and nothing is working to stop the error.
Your crash is related to Radeon settings.
1.Disable Radeon settings from auto start(ctrl+shift+escape>auto start tab)
2.Completely uninstall Radeon settings(custom uninstall) or reinstall drivers without Radeon settings(custom install)
3.You need just try this steps.
4.Full model name of your GPU?
The black screen seems to happen when Radeon settings load. I was not aware you could disable the settings from loading, I will look into that.
XFX AMD Radeon R9 390 8GB GDDR5
DD Radeon R9 390 1015MHZ 8GB D5 DP HDMI
Part No R9-390P-F28M
I did a custom install of the latest drivers to not include Radeon Settings. After a few minutes the GPU crashed to black screen anyway.
I tried MSI Afterburner again. Instead of 950mhz though, I had to change to the lowest setting of 750mhz to prevent all issues.
Regardless, I can see the Mem Clock jumping constantly from 150mhz to 1500mhz every few seconds, sometimes triggered by moving the mouse or dragging a window, etc. Something is wrong with this GPU and since it's happening to many others in this thread and related, there is clearly some kind of error with this card. Since it is now happening on all drivers it seems to be a universal malfunction that needs to be addressed.
My old GPU that is working fine with catalyst drivers has 1375mhz clock, so dropping my current one down to 950 or lower means there is no point in keeping this GPU.
Nope, even at 750mhz the problem happened again.
Wattman does not work with 390 cards (imo, never will) . Ab, ocguru, any oc /optimization app will cause problems (with 12.1/2/17.1 for many(imo).
Delete them all, run card at stock, or go back to 16.11.5..
And Clean Install AMD Graphics Drivers (now need newest ddu for relive), has been mostly successful for faulty installs.
Old cnext probs were mostly related to faulty/corrupt/missing, msvcr dll's(win os ?), yours is ntd.dll(can't offer fix for that one), from the one you supplied (how many others ?)
May be your card is starting to fail, but if you go for RMA, you have to uninstall all unnecessary software(not games) ,to prove it.
I've never used (or needed ab, etc..) to run any of my cards (and I oc them at some point, not before..).
I'm running a z77 board (newer (2016 beta) bios), R9 390G1 with the newer ?(2016) vbios. win10 pro.
Go back to basics for troubleshooting (one 3rd party app could be the bug).
I've uninstalled my drivers completley several times using those steps. I went deep into windows and made sure nothing is being downloaded automatically. I have downloaded several different drivers in several different ways and the problem always comes back.
What is wattman? I don't use any optimization apps. Only MSI Afterburner, and only as an attempted solution.
Cnext probs seem related to driver but I have installed multiple different drivers so that's not the issue.
Not likely any 3rd party. I haven't added any 3rd party recently and everything was working fine and works fine with other GPU.
You mention vbios as a different thing. Since i had to update my BIOS to get this GPU to work in the first place, the issue seems related to BIOS. How to I check/change/fix vbios? All I know is that I have the most up to date BIOS settings for my motherboard (Jan 2016)
1 of 1 people found this helpful
Wattman replaced Overdrive in the newer crimson relive drivers.
Vbios is for graphics card, contact xfx support (some manufacturers list one in download, others will/may offer one through their tech support).
With all the stuff you've tried, comes down to OS problem, or card fail. May be RMA the card time.
Sorry, can't think of anything else.
I does look like others users who have had this error were able to correct it with their RAM settings. You could try selecting auto for the RAM profile in BIOS and see if that fixes your issue. If you don't see the black screen, you could then try reverting to an overclocked profile for the RAM if there is one or if you are already in AUTO.
people said they tried this and it didn't work. The fix seems to be down-clocking the memory which is lame. I'm looking into XFX support now.
I checked my BIOS and my RAM settings have been set to AUTO the whole time.
I installed crimson 16.11 and am using MSI afterburner to limit memory clock to 1400MHz. This is somewhat stable but not a cure. The memory clock is jumping to 750MHz constantly for no reason. The problem returns but not as frequent.
Now that my GPU is somewhat stable by limiting the memory clock, I tested some video games. In each case the screen flickers black a few times and after about 5 minutes the game crashes with some kind of error.
Same problem here also. I created an account here specifically to reply to this topic. I was on the latest WHQL Certified 16x driver, and when I tried to update to the "optional" non WHQL Certified driver version 17, my screen goes black towards the last 15-20% of the update. I tryed updating through the crimson software, and also by direct download from AMD. EVERY TIME I try to update to the optional update it crashes my GFX card output.
MSI H81-E33 MoBo
Intel i7 4790 3.6-4.0 Hyperthreaded Quad
Windows 10 Pro x64
16GB Crucial Ballistic Sport (2x8 1600 DDR3)
Asus Strix 390x Super OC DCU3 Edition (RUNNING @ FACTORY WHEN THIS OCCURED)
Rosewill RBR 1000W Bronze PSU
Acer x223w monitor connected in this fashion:
DVI to HDMI Adapter attached to the DVI port on the 390x.
HDMI Cable to another Adapter on the monitor.
DVI to HDMI adapter plugged into DVI port of monitor.
(DVI ( HDMI)-----------------(HDMI ) DVI )
So I usually held down the power button and hard reset the PC once I realize the output signal is never coming back, this time I let the PC run until I was sure the install completed and my hard disks stopped making noise before I hard reset; now I'm stuck at a blue screen with the windows 10 loading dots spinning, and below it says "Just a moment". Well this moment has lasted 20 minutes... Looks like I'll have to wait for the next next official WHQL Certified driver... My AMD Gaming Evolved program refuses to stay running also. After it loads up and I get a chance to open the log in window, it completely crashes. I don't know who to blame, Microshaft or AMD, but the new version 17 driver is causing severe damage to users operating systems, and causing me a lot of wasted time. Let this be a lesson to all of you, make sure you create frequent "restore points" for your system to revert to, if it wasn't for that I'd be seriously surprise fornik8ed in the posterior. And a note to all, I'm a highly experienced user, so don't even bother mentioning ridiculous suggestions like "BIOS updates" and such to me. My system receives over 15 hours of weekly maintenance routinely as I have an obligation as a server. My system was running like an Olympic gold medalist before this latest driver, now it's as spry as Christopher Reeves in his later days...
2.Leave only one monitor with native cable connection (no adapters)
3.Delete C:\AMD folder
5.Then reboot and plug your remained monitors one by one,checking at each step, that win10 has installed correct driver for each monitor
2 of 3 people found this helpful
Anything that can be updated, is updated. I tested the driver in my test-bed which is a fully updated and "VIRGIN" win10 x64, as in not one single program installed or setting changed with the exception of official Microsoft Updates. The virgin gets wiped after every use and is the "control" for all my software troubleshooting. There is an issue with Microshaft or AMD. I am far from the only person with this exact problem. To anyone reading this topic:
If you have a Radeon 200 or 300 series card, and Windows 10 Pro 64 bit, and when you try to update to the latest driver you get a black screen, it's AMD's fault. They don't care about you or or your problems, and they have CONFIRMED that this issue will not be fixed publicly. This is either a BS tactic to back us into a corner and purchase new hardware, or just a giant **** you to their customers. Either way, this 390x will be my last AMD product. Good luck to all and hope you guys find a solution. My solution was a GTX 1080ti.
I am only now noticed your terrible monitor connection system
First of all you must buy a good DVI-DVI dual link cable and connect your monitor without any adapters.
And tell me, why did you created such monstrosity?
This is definitely a huge issue that needs to be addressed still.
Happening to me also as of 8/20/2017. Can no longer boot windows to native resolution of monitors with AMD drivers. As soon as we choose above 1280x1024 black screen, monitor is in standby, machine is frozen requiring hard restart.
There are no relevant events posted to eventlogger when this happens.
Systematically eliminated some other causes...
- switched GPU to Radeon HD 4850 (which was working fine in another rig) and problem continued (eliminating the r9 hardware)
- replaced motherboard and updated to latest BIOS
- removed all unnecessary peripherals, hard drives, PCI. PSU is 750W Corsair. Power consumption from UPS shows max at 295W
- updated windows 10 to latest, then to fall creators beta, then fresh installed windows 10 (still happening!? this is now baffling... did GPU get fried?)
- booting to safe mode or removing all AMD drivers with DDU leaving only Windows default driver for video; issue goes away. tested with 2650x1600 (proved all the hardware is still okay)
- using DDU to remove as instructed above and adding older 16.11.5 ... problem returns (something is now buried in Windows causing all drivers to fail, persisting through a Win10 OS refresh and DDU)
- using MSI afterburner to underclock Core to 950MHz does not work but underclocking Memory to 1200MHz has been stable for a couple days now.
update: tried latest AMD driver v 23.20.780.3 from 2017-10-11 and no change in behavior.
We are stuck underclocking memory to 1200MHz until this issue gets figured out. Also, should add that multiple monitors still cause the issue regardless of underclocking.
SOMEONE PLEASE FIGURE THIS OUT!!!
Please do not resurrect a 10 month old thread, open a new one instead.