Graphics Cards

[redcow16]

i have been getting BSD for a while and im tired of it. it always happens when i game, it doesnt matter what game i play.

i recently found out that the omly profils in adrenaline that works are the default and the quiet mode. balanced, rage and customs are always giving me BSD.

exemple: im playing overwatch, temps are fine, gets a blackscreen, entire PC reboots by itself with no video output and the debug LED lit for the GPU. i turn off my pc by holding the on button, let it reboot, log back in to a wattman notification that says : default radeon wattman settings have been restored due to an unexpected system failure.

is it just AMD drivers and adrenaline that is dog**bleep** or is it a hardware issue whit my GPU?

if anyone knows how to fix it  let me know

PSA. temperature is not an issue

specs:

R9 5900x

Asus X570 crosshair hero VIII wifi

32GB corsair vengeance 3200Mhz CL16 (8x4)

Asus TUF RX6800xt

Corsair RMx 850w

[redcow16]

the default preset is also giving me BSD now smh...

[riversde]

Did you use AMD cleaner or DDU on safe mode first?

[Harry3699]

Ok I wrote a post about my experience of black screening with Radeon cards, short version tried DDU uninstalling etc but no good, problem was PSU. NOT related to power of PSU I replaced a 700w PSU with an older 650w PSU differences were that the one that worked was Dual rail not single rail 12v, had 179w on 3.5 5v rail instead of 150 and was semi modular against non modular. I am trying to get another PSU to track down which single or combination of these is the cause. Personally the dual rail seems to be the most likely reason but yet to be tested. Black screening seems to be a real problem particularly with Sapphire cards and if all the usual driver fixes don't work try this it worked for me.

[redcow16]

DDU safe mode always

[ThunderBeaver]

My build.

MOBO X370 M7 ACK

CPU R7 2700x at 4.2GHz

GPU RX 6900 XT VPU at 2200MHz and VRAM at 2170MHz.

RAM Corsair Dominator Platinum 32GB Dual Channel PC3200

SSD Samsung 870 EVO SATA 3 4 TB

OS Windows 10 Pro 64 

PSU Thermaltake Toughpower fully modular 1200 watt.

Every one I've heard dealing with black screen issues on the RX 6800 XT and 6900 XT are using PSU's under 1000 watts. The specs I've read on these cards is their base power draw is 300 watts. With overclocking they can pull anywhere from 450 to 525 watts max. Your system may be protecting itself from power surges and/or hardware damage due to inadequate PSU. The power supply I mention is single rail and only costs about $250 on Newegg.com. Even AMD recommends using at least an 850 watt PSU. With all the overclocking I'm doing I needed the headroom so that's why I got the 1200 watt PSU. Haven't had a single issue and I don't use adrenaline overclock presets. I custom tune my cards and stress test them before gaming. If no artifacts or tearing occurs I keep pushing until I see that. Then I step the clock timers back about 50 to 100 MHz on both VPU and VRAM and haven't had any issues yet. Hopefully this advice helps in your search for a solution. Good luck and nice setup.  

[redcow16]

i have an 850w psu running stock everything for my gpu and it still crashes when i game but when i do benchmarks like heaven 4.0 and 3dmark its never crashes

[Geforsikus_2021]

There may be problems in the video card, the power supply should be enough with a margin (if considering only games, then your processor in gaming loads will consume up to 140 watts + the video card 269 watts in peak in the amount of a maximum of 500-600 watts), I would check this video card on another PC. Alternatively, there is a wonderful BlueScreenView program for reading blue screens. Either the video card is faulty, or the power supply gives a drawdown on the 12v line (and the card is cut off due to a lack of power supply), although in theory a famous brand, and the power supply is not cheap, there should be no problems with changing profiles. It would be better to check both the power supply and the video card on another PC before buying, if possible. Run another stress test of the CPU, and RAM just in case, first everything separately, then together.

[redcow16]

i already bought a new psu, ill see if it still happens and ill also run memtest86. i dont think i need to stress test my cpu again because it passed A1 cinebench R23 stability test which is running in a loop for 30min

[Geforsikus_2021]

There were problems with Seasonic power supplies, it was an expensive line of units, they cut off the power supply because of smart protection when a strong load was applied sharply, and the power supply went into protection allegedly due to overload. Later they fixed this jamb. Korsair didn't seem to have such problems. You can monitor the voltage in the aida64 program (computer tab, sensors. ) without load and with load.

[redcow16]

i ordered myself an AX1600i from corsair. we'll see where that gets me if it fixes it or not.

[ThreeDee]

One thing is to make sure you are running separate power leads to each power input on your GPU

Make sure your motherboards BIOS is up to date as well

[redcow16]

i never run pigtails for gpu power and since i got the AX1600i i can monitor it through iCue and if anything ill use one of those aida64, HWinfo64, HWmonitor, etc...

[amdman]

"i recently found out that the only profiles in adrenaline that works are the default and the quiet mode. balanced, rage and customs are always giving me BSD."

I would place a bug report to AMD, I only use the standard default with no overclocking one and turn on Smart Access Memory and never had a problem with my 6700 XT, so it seems to point that there is a problem with the profiles. I hope they can fix it.