Hello,
i have a problem with dx12 and vulkan games on RX6800. In games using these api appear random black screens. Games running on dx11 and dx9 works perfect. I tested on Star Wars Battlefront II and on dx12 mode this game causes black screen but when i switch this game to dx11 game works stable. Serious Sam 4 on vulkan causes black screens but on dx11 work stable. Cyberpunk 2077 on dx12 causes black screen. I tested some dx11 and dx9 games like Realm Royale, Paladins and Guild Wars 2 and all those games work stable. Black screen dont appear on gpu stress test, and desktop activities like web browsing.
If black screen appear pc dont respond but i hear sound few seconds, and i have to push power button for restart.
PC specs:
GPU: RX6800 Saphhire Reference
CPU: Ryzen 7 5800x
Mobo: MSI B450-A PRO MAX
Storage: Samsung 970 Evo NVMe PCIe M.2 500GB
RAM: Corsair Vengeance LPX DDR4 3000 C15 2x16GB
PSU: be quiet! Straight Power 11 750W
Radeon driver: 21.6.2, tested most older versions and same problem. On 20.12.1 driver dx11 games causes black screen too, but on newest drivers only dx12 and vulkan.
Anyone know some solusion for this problem?
I recommend replacing the motherboard(many people have this board burned), given your processor, perhaps the power supply chains can not cope with the load, you can try to update the BIOS to the latest version. Or if you find out exactly what is wrong, connect the video card to another PC, and install firewood there and try to play. If there are no problems, then the motherboard does not pull 5800x, the maximum that it was originally designed for is the 3000 series of processors, despite the fact that support for the 5000 series has been added, 5600x may pull. Directx 12 loads 6-12 core processors perfectly, and directx 11 loads 1-3 cores poorly.
Thanks for answer. I tested on a Ryzen 1600 processor before and had the same problem. I also tested a different graphics card RX 570 8GB and it worked stably on the Ryzen 1600 and Ryzen 5800X. At the moment, there's the latest bios on the motherboard. 5800X stress test is stable. Im tested RAM too, and 0 errors. I don't know if it's important but monitor is Xiaomi Mi Monitor 3440x1440, 144hz via displayport.
This is definitely not RAM (the system would be unstable and the PC would turn off or reboot), if the rx6800 on another PC causes the same problems, most likely the problem is in the video card itself. I correctly understood that during the game on directx 12, the game itself works but there is a black screen on the monitor(there is no inscription on the black screen, for example: no displayport signal? Alternatively, try connecting either another monitor or a TV if it supports the display port. Or reduce the refresh rate of the screen to 120hz. Perhaps the displayport version of the cable is 1.2, and you need 1.3 and higher 1.4 1.4 A.
I tested the 120hz refresh rate and after 3 hours of play a black screen appeared. Previously, the problem was between 5 and 30 minutes of play, but in this case it could be just luck. When there is a black screen, the game works for a few seconds and then the computer shuts down, but the fans are still working until manually turned PC off with the power button. The word "no singal" appears on the monitor and monitor goes to sleep state. Windows event log just say unexpected shut down, Event ID 6008.
Did I understand correctly, the rx570 card behaves stably both Ryzen 1600 and 5800, and the rx6800 is not stable in both cases on 12 directx? Tell me which cooler or liquid cooling system do you use? And In the stress test of the processor, how does it behave, what temperatures and frequency does it keep in the stress test? Try using amd adrenaline to fix the fps below the monitor frequency for example: monitor 144hz and fps 141 or 120hz fps 117(freesync can also be turned on) And download the CrystalDiskInfo utility to check the status of hard drives. Has there been such a problem before, or has it just not appeared for a long time?
RX570 is stable in any case, but performance is low. I have had the RX6800 since the premiere, I bought it to play cyberpunk 2077 and there was a crash every few hours. I thought it was the fault of the driver because it was the first version of the driver. After that, I only played dx11 games for half a year and there were no problems, but when I started Star Wars Battlefront II on dx12 yesterday the problem came back. It's very strange. I am not able to understand how dx12 games cause crashes on this graphics card. The problem occurs with both freesync on and off. Enabled and disabled resize bar but only in dx12 and vulkan games. CrystalDiskInfo says the disk is 100% good. Maybe the fact that the motherboard has pcie 3.0 and the graphics card 4.0 is a problem? I have this cooling system https://pl.aliexpress.com/i/4000269208885.html , not very good but max temp in stress test is 80 celcius, in gaming 70 is max.
The RX570 is stable in any case, but performance is low. I have had the RX6800 since the premiere, I bought it to play cyberpunk 2077 and there was a crash every few hours. I thought it was the fault of the driver because it was the first version of the driver, and cyberpunk was very bugged game. After i finish cyberpunk 2077, I only played dx11 games for half a year and there were no problems, but when I started Star Wars Battlefront II on dx12 yesterday the problem came back. It's very strange. I am not able to understand how dx12 games cause crashes on this graphics card. The problem occurs with both freesync on and off. CrystalDiskInfo says the disk is 100% good. Maybe the fact that the motherboard has pcie 3.0 and the graphics card 4.0 is a problem? CPU cooler is Snowman MT-6, on stress test CPU had max temp 82C, most time was 75C, and frequency was near 4.6Ghz on all cores, tested one hour. GPU have sometimes 92C on hotspot but crashes are random not only on high temperatures. Crashes was also in one dx11 game The Beast Inside but whats interesting only in one cutscene always at the same moment. That was the same crash like in dx12 - blackscreen then pc shutdown, but in this game problem gone after some amd driver update and now this game is stable.
Strange behavior of the video card, there are articles about what amd has acknowledged the existence of problems with a black screen with the subsequent shutdown of the PC. AMD Radeon RX 5700 XT (Navi 10)
AMD Radeon RX 5700 (Navi 10)
AMD Radeon RX 5600 XT (Navi 10)
AMD Radeon RX 5500 XT 8 GB (Navi 14)
AMD Radeon RX 5500 XT 4GB (Navi 14) Probably rx6800 and also got under the hand of this driver bug or defect of the graphics card, Or wait for a driver update, if there is a guarantee on the graphics card try to change rx6800xt surcharge, tembolee that prices have streamline fall on the graphics card.
Now i see one more thing. When crash is occur on mobo vga led is on until i restart pc on power button. Documentation says that means "GPU is not detected or fail".
Try to play with the frequencies of the video card again, through amd adrenaline (the performance tab, manual configuration, manual configuration management, the GP configuration check box and the advanced management check box. do not touch the voltage, but reduce the minimum and maximum frequency, you can set the same value, only the maximum value will be set in the drain by default, reduce for example: (2400 mhz by 100 mhz to 2300 at least) and apply, you may have a different value)
Perhaps, it is only possible that the problem may be not only in the driver for the video card, but in some update on windows 10. The behavior of the video card, as if it is removed or disappears from the system, as sometimes happens with hdd or usb flash drives. If there is such an opportunity to check the video card on another motherboard? Return the default settings of the amd adrenaline drivers. And turn off all the checkboxes in the video card tab, except that the frame rate control is set less than the monitor refresh rate. If nothing helps, then either wait for a new driver or hand over the card under warranty.(
I reinstalled windows but still crushes, lower clock speed, and fps limit dont work too. I will check this gpu on another pc in future because i cant at the moment. Thank you so much for all your help!
Try the driver version 20.12.1 they write that there are no problems with a black screen in it, or this is a matter of a specific case) Watch the video on youtube (Ancient Gameplays) he makes comparisons of amd drivers there and read the comments, very interesting, maybe you will find a person with a similar problem as you have and ask what driver he uses)
I tried this driver, and that was worst driver for me. On this driver i had crashes even on dx11, probably because of this issue "Some displays such as the Sceptre C series or Samsung™ Odyssey G9 series may experience an intermittent black screen on Radeon RX 6000 series graphics products.". In next version 21.1.1 their repair this issue and from this version all dx 11 games work stably.
Well, then the problem is most likely the driver, if everything is worse with the old one, then it either wasn't fixed or didn't even try, all the options in the driver settings are disabled?
All options are disabled, i tried minimal install without adrenalin software too. I will problaby check this gpu on other pc tomorrow.
I tested it on other pc, and identical problem. This gpu is probabli broken. Both pc have pcie 3.0, it is posible this dont work because not pcie 4.0?
No, it is unlikely that it is broken, it's just that amd still can't fix this bug, despite the fact that some drivers seem to have fixes related to the black screen. Take the behavior of the video card on the smartphone camera, and hand over the video card under warranty, and take nvidia. Pci express 3 and 4 are backward compatible, they just have different data transfer speeds, the difference in tests is extremely small(1-2fps) Perhaps this is just my assumption that this problem occurs if both the processor and the video card are from amd, but these are just thoughts out loud, since there are mainly problems with this hardware configuration.
If this goes on, Amd will simply go broke on video cards as they will be constantly changed because of this problem. Judging by the statistics of the steam pack, everything is bad for AMD. I myself recently switched to rx6900xt, before that there was a gtx1080ti, there are no problems with the pack, the only thing in some games is the performance of the card on 11directx is extremely bad, the card works from 60 to 80%, but I think this is a problem of the api itself, since rtx 3090 behave exactly the same. There are no problems in games on directx 12 and vulcan, the card works at 99% if the fps is not limited, so for now I am more than satisfied with the video card.
Yesterday I was at a friend with my graphics card who has a similar PC with the difference that he has a ryzen 5600x and an MSI B550-A PRO motherboard. My rx6800 work stably without a crash for 2 hours, in a game that I mostly crash after a few minutes. Perhaps I should buy this motherboard.
Yes, this will most likely solve the problem, but given your 5800x processor, I would recommend a higher-class board on an X chipset. If it doesn't bother you, try checking your 5800x processor with a friend, too. To make sure that the problem is exactly in the motherboard(or in the processor). And you don't have AMD Smart Access Memory enabled? Since it can cause problems with a black screen. Just if you buy a new motherboard and it turns out that your problem will not solve, then the problem is in the processor(
Alternatively, turn on the power supply circuit (high performance. And not (amd ryzen balanced) So that the processor cores do not park during downtime, so that they constantly receive voltage.
I think the motherboard power section is poor. If I do a CPU or GPU stress test it always passes well, but if I do a CPU and GPU stress test at the same time the PC crashes sometimes. This is probably not the fault of the power supply because my friend has the same power supply model but 650W, and I have 750W.
Everything is possible, but to make sure, check your processor on a friend's PC, or just buy another motherboard, and you have a good power supply, perhaps the vrm on the motherboard is overheating, although it is strange that there are no problems only under CPU load. Load everything together and look through the hwinfo64 utility for data on the voltage and temperature of the CPU. And here's another video card, connect it with two different cables from the power supply, there can be a divider for 2 po(8pin) on one power cable. 2 cables should go from the PSU to the GPU(you have). I have 3 cables from the psu to the gpu)
I ordered a MSI MAG B550 TOMAHAWK motherboard, I will test it when it arrives. People think that it is good for ryzen 5800X. Today I tested the game with a detached sound card and game was stable for 2h without crash. It's very strange.
It is really strange, in some cases it helps to simply pull out the RAM and processor and re-install (somewhere the contact is bad) in different ways, the MB driver on the sound card is buggy or the high audio sampling rate is set by default(on an external sound card). It helps someone to simply flash the card with a new version of the BIOS, since in the first revision there are problems that, according to AMD, have already been solved in the latest drivers.
I replaced mobo few days ago, and still the same problem. Yesterday i noticed even bluescreen, and artifacts in dx11 game. I will contact the seller about the warranty.
How do you have a video card connected, 2 cables go to it from the power supply?(take a photo and attach to the message how the wires from the power supply unit go) There should be 2, one cable holds 150 watts, two of course 300 watts. If one connector is connected, there may be problems under load. Just if they take the card under warranty and check it and they do not have a problem, they can return the card to you back. I have an rx6900xt non-interference, 3 8 pin additional power, 3 separate cables come from the power supply (450 watts in total will withstand) But the card consumes less. see the screenshot. You should have 2 cables running. my power supply is as old as a mammoth's poop)))) but he is already reliable for about 10 years +-)
For me, 2 cables were connected on each tested computer. I wrote to them that problems can occur even after a few hours of playing, and are generally quite rare.
A really interesting case for you, try another pack card you have to download the directx runtime offline installer package, in the first link on the microsoft website it will give out in the search bar, it weighs 97mb, unpack and update it and also check whether microsoft visual c++ packages are installed on your pc if not, then also in the search bar microsoft visual c++ all in one there are all versions of the bundle, it is unlikely that it will help, but it's worth a try, ask in the amd sapphire radeon community, maybe fresh bios for your video card were released or not, maybe the problem is in the first bios revisions. Of course, you can still take down the Windows, but considering that another RX 570 8GB video card behaves well in this system, I think the problem may not be in it, try to move the cables from the power supply to another socket in the power supply that power the video card (the power supply itself is excellent). You can also try using the Trixx utility( from sapphire) to switch the bios of the video card, but most likely it is one in your model and it will not work. Sorry for the crooked English
Im tried to move cables to another socket in PSU, reinstall windows, and use trixx (my card have one bios). I can only wait at the moment for results of warranty. I send this card yesterday. Thank you for your help.
I hope there will be no problems with the exchange of the video card, good luck!)
ansuu, Please let us know how the warranty return goes. I'm having this same issue with my ASUS Tuf 6900XT.
Black screen on certain games after random periods of playing. After reading about the issues people are having with the New World game, I'm sort of wondering if during certain periods of games the card ramps up fast with a spike, before it levels out. The Voltage regulator trips and shuts device off(black screen with working machine.) NVidia recently had to release a patch that reduced card speed and reduce voltage to help with this. You can see more on this insight on JayZTwoCents Youtube.
I have noticed that I go to black screen on game loading/high gpu usage; i.e Warzone when you zoom in to the heli's dropping you off. I've also noticed issues with Wolfenstein II crashing.
I've tried all the driver updates and downgrades to earlier versions. Reinstalled Windows. Removed mismatched sticks of ram down to 1 stick. I'm on an i7 7700k with Z170 motherboard. Just wanted to throw my hat in the ring with my issue on the forum.
Its probably not overvoltage problem, im tried undervolting and reduce clock speed to 1500mhz, and gpu was still the same unstable.
Some time ago, there were problems with the seasonic power supplies, due to a sharp fast load on the video card, the power supply thought that there was a short circuit and cut everything off, the power supply's circuitry was too sensitive. If you have not yet returned the card under warranty, try powering it with another power supply.
I found this thread https://www.overclock.net/threads/psa-how-i-fixed-my-black-screen-issue-and-further-increased-stabil...
It talks about on older Intel motherboards, instability with lower voltage on your VCCSA and VCCIO values within your BIOS. They recommend upping to 1.2/1.3V. I have an ASUS Z170 Pro on a i7 7700k, and on XMP profile, my VCCIO was already 1.3. My VCCSA was 1.2. I went ahead and upped it to 1.3V.
I did this change and so far I have been able to play Warzone without any issues for a few nights. I'm still testing but wanted to give a preliminary partial thumbs up of it fixed my problem. I know that doesn't help you ansuu, but could be an indication of weak memory controllers needing more voltage to support the bandwidth of the video card.
check the VRM heatsinks,, they use some sponge washer make sure they are tight enough not like HOLD A BUILDING tight but snug.. wiggle check to make sure.
Hi,
that's the end of the story. The graphics card went to the service with the description of the fault in short "causes blackscreens in games based on api vulkan and dx 12, but works in dx 11 games". The service staff proved to be incompetent because they tested the card for only 10 minutes in a program using dx 11, and returned it to the seller as working condition. Therefore, I do not know if the graphics card is functional or not because the test procedure was idiotic. Fortunately, the seller agreed to refund me for it and today I ordered the rtx 3070.
Thank you very much for all your help!
Sorry to hear that. I had to send my card into ASUS as I tested it in another machine with newer hardware and it crashed within 5 minutes of running 3DMark. Hopefully they find fault in the card and send me a new one that works. No amount of config changes was able to make the card stable in gaming/3D testing. Will update when I get a return.