So, I keep getting a black screen on my rx 5600 xt, and I think a lot of people have the same problem. I'm having this issue since I've got this card. It's really frustrating and I've lost some work because of this, since i have to turn off my PSU, turn it back on, and power on my PC. I've been trying to fix this issue since the problem began, but I didn't succeed. I would really appreciate if someone could help me, because I really don't want to send the card back, because it's it's a great card, except for the black screen issues.
MoBo: Asus Prime H-370 a
CPU: Intel Core i5 8600k
GPU: ASRock AMD Radeon RX 5600 XT Phantom gaming 6GB OC edition. Factory bios.
RAM: 2x8 GB Corsair Vengeance LPX DDR4 2400 MHz
Storage: Samsung SSD 850 EVO 500GB, Samsung HD103UJ 1TB
PSU: 550 watt.
Radeon Software is overclocking these 5000 series cards out of the box.
This results in higher temps, more fan noise and for some cards: instability and crashes.
Look up the boost clocks for your make and model from manufacturer spec sheet.
Then enable manual performance tuning within Radeon Software and enter your
max boost clock into the "P3" power state. Then hit "Apply Changes".
Das hat mir sehr geholfen !
Die neue Software 21.2.2. hat diesbezügliche Probleme mit meiner RX5700XT beim Gamen gelöst.
Stelle in Win10 bei Einstellungen, System und dann auf der Seite unter Grafikeinstellungen von Windows auf AMD Software um, ich habe hohe Leistung eingestellt!
I keep getting a black screen. I start up my PC, I open the browser, black screen. I restart my PC, I open the browser, black screen. It took me 3 tries to get to this post. And I will happily switch back to Nvidia, because my old gtx 680 from 2012 is still better for gaming than an rx 5600 xt, because I DON'T GET A BLACK SCREEN! If anyone who works at AMD is reading this, I will send my card back, because if I'm honest this card is a piece of crap. I may sound rude, but these issues are going on for far too long.
And apparently I will have to send back my Oculus Quest 2 that I just ordered, cuz my GPU in now too weak. Thank you AMD.
And if there is one issue that is going on for like 2 years already, may I ask, how can the name of your company begin with ''Advanced''?
Well it's the brand of card you got. I had a gigabyte RX 5600 XT and it worked fantastic for the 7 months I had it. The black screen can be a driver issue related to using one past the 20.8.3. If you care to try it out, run DDU in safe mode, with the internet disconnected. Then install 20.8.3, disable the auto update feature in Radeon settings and you should be good minus any other settings you mess with.
AMD Previous drivers, scroll down to the 20.8.3 and save it: AMD Radeon™ RX 5600 XT Previous Drivers | AMD
The latest driver has something in the release notes about a VR fix that might interest you.
Your GPU is plenty to run the device: Oculus Quest 2 system requirements | Can I Run Oculus Quest 2 (systemrequirementslab.com)
How they get "Advanced" not sure but they are "advanced" and I can say it's a niche community, "fan boy" thing to own an AMD anything, since they're not always click n go, like some others but yet consistently outperform the competition when properly setup.
To get into the settings that @RPX100 mentioned, you must go into the "performance" area of the Radeon software and click "manual" accept the warning and continue. Enable all the disabled items, won't hurt a thing in there, leave VRAM alone and there's nothing to really mess up in there. Let's see if the card gets stable first.
@mackbolan777 what AMD is doing with it's driver + Radeon sw is bonkers. You can never know the cause of your crashes... I've had issues with my RX 5600 XT since I bought the **bleep** thing. I tried everything, every driver version since the release of the card. DDUd the drivers, reinstalled Windows 10 several times, but no resolution for these issues. Meanwhile, when I was getting sick of these issues, I would just plug in the old RX 580, and it would work without any issues. I even upgraded my PSU, thinking that my 5 months old Seasonic 550W could be too weak for this card paired with a i9 9900K.
What I eventually did, was this: I DDUd the driver, and let Windows install whatever it wanted for the card. Windows installed a driver from August 2020, with no Radeon software. That happened 4 days ago, and I haven't had any crashes since. Maybe they will reappear at some point, but for now it works. I have to say though, that I did have the AMD driver from August 2020, and I got a load of crashes with it. Because they bundle the driver with the crappy software, it's very hard to know the cause of the crash. Sometimes the Radeon software would crash itself when trying to start it. This Radeon software + driver issue is like having a possibly drunk driver, driving a possibly broken car. Because you cannot root out any of the two causes, you never know which one is at fault.
You can do what you did by manually installing the driver pack and stopping it at he part where it shows the install directory. By that point the driver is extracted to the specified path in the window. Copy or write down the path and cancel the install. Go into device manager find the display adapter, click "update driver", "browse my computer", and navigate to that location that was in the installation window.
You've now installed just the driver, no Adrenaline and it is a fix that has been tried. As I mentioned, the 20.8.3 (August 3, 2020) is the most stable driver for the 5000 series in general. Rumors are the newest driver works for some but I wouldn't bet on it. The 5000 cards are EOL (End Of Life) so no new fixes for the most part.
A whole other process, that albeit shouldn't need to be done by an end user, is to go into the tuning setting to set some things up. Click on "automatic" in the OC area, select "undervolt GPU" and write the number down. Switch to "manual OC" enable all disabled things, in "fine tuning" for the voltage, put that lower number you got in there as the "max" voltage. Slide you core clocks closer so the 1st clock is ~1000, middle ~1300, max ~1650 (your clocks may be different but that's the idea. Should be ~960mv. I select the option to to disable the "zeroRPM" fan function as well. Finally slide the "power limit" to max. This allows the GPU to add voltage under load if needed. Hit apply. Save the profile in the upper right by clicking the "gear" icon, "save profile".
Try those settings out and see if any problems arise. If you crash again, reapply the saved profile and add +12Mhz to the VRAM, try again.
Sometimes the crashes could be RAM timing or bad RAM, so I would run MemTest Manual (hcidesign.com) to 800% coverage to be certain there's no issues. The CPU OC, if applied and "uncore" is not enabled, I know we're talking Intel, it could create an issue because it will cause the BUS speed to change, altering the PCIe multiplier.
Then there is brand of card. Not all "partner" cards are created equal. I had a Gigabyte RX 5600 XT (2 fan) and it ran fantastic. Always run 2 x PCIe power cables if you have 2 x 8 pin connectors, it makes a difference.
In 2018 I had an Intel 8700K paired with an MSI Duke 1080i and had WHEA 18's from time to time, crashes and it was BIOS settings, while running the click n go 5Ghz option and Nvidia had a problem driver as well. So not always an AMD driver issue, but it does catch the blame due to several bad experiences.
It's easy to blame the part or driver but comparing an RX 580 to the RX 5600 XT is not a good idea, they use different aspects of the driver package. They're drastically different in design as well.
@mackbolan777 I tried everything, from undervolting to underclocking, but I still had crashes. My RAM is at default settings, no XMP, 2666MHz Samsung RAM. The CPU is at stock 95W TDP, no multicore enhancements. I tried installing the AMD driver by decompressing the installer using 7zip. It still crashed. I ran MemTest for 24 hours without any errors. The only thing that initially reduced the crashes was writing the old 12000MHz vbios on the card. And now with what's installed by Windows I have yet to run into a crash.
I doubt 5000 series is EOL, since even the 500 and 400 series are still receiving bugfixes and improvements.
I have a solution I have posted a detailed reply on "ackerworx___dot__UK" but I will try and reexplain it in a textual format here, but its dependent on a particular error, check your eventlogs and look for any Dcom errors,
*** Verification Step - You Can Skip this, but it's better to make sure****
If you find them, clear the log and try and replicate the crash. Then check the log for another similar Dcom issue.
*** Verification Step - You Can Skip this, but it's better to make sure****
On Windows it's EventLog ---> Windows Logs ---> System
Go to start ---> type "Component Services"
In Component Services in the left hand panel showing the Console Root tree
Click on the plus sign to expand the branch that says "Component Services"
and click on the next plus sign beneath that says "Computers"
In the Middle panel/pane right click on "My Computer" to open up the Context Menu and select properties
In the "My Computers Properties" Window, select the tab "Default Properties"
Unselect Enable Com Internet Services on this Computer
Unselect Enable Distributed Com Services on this computer
Click OK and restart
PS: This does NOT disable your internet, DCOM and COM is part of an old archaic system, that is rarely used in a home computer environment.
If your in a server/client environment, check with your system administrator first to see if disabling this will affect any services or legacy services that you might connect to.
EOL, you're partially right. They changed course I just read and they will continue until Q1 this year. The 5700 is EOL now. The RX 5700 XT and RX 5600 XT are winding up in Q1 2021. However, I'm not certain on "fixes" they offer in newer drivers. So far it's been nothing new since 20.9.2 for any of the cards you mention.
You mention the vBIOS, some cards couldn't handle that and an RMA should've been done in those cases. That was a "maybe it will be stable" idea that did work on some brands. Gigabyte was one. Others were hit and miss. So if you tried to write the vBIOS back, check in GPU-Z to see if it wrote back, I was told that's a one way street. In that case, your card should still be under warranty, RMA it.
@mackbolan777 It did write the vBIOS ok. I checked with GPUz and with Radeon software. The memory was at 12GHz on both counts. About RMAing the card, I tried that as well. The problem is, the ill-informed people at the warranty service don't understand that you can't test a card just with Heaven. That's all they did, and sent the card back with "defect not manifesting", even though I've sent them videos and screenshots of the card misbehaving. Another problem is that with the current prices, even if I would get my money back, I could barely get a new 1050 ti with what I paid for this 5600 XT, which is not exactly what I want right now for 1440p gaming
Hint, sell it. Mine worked fine and sold for $500 on eBay. I'd try it out in a friend's machine first to make sure it works. Then sell it, send a redacted copy of your receipt/invoice with it, just in case the buyer needs warranty. I agree, that is total BS that the manufacturer ran one test and called it "good" or even bothered sending the exact card back. They should've sent you the properly revised card with the updated vBIOS or PCB. However, tough times, you can easily make enough off that to offset something better.