cancel
Showing results for 
Search instead for 
Did you mean: 

General Discussions

sypei
Journeyman III

rx 580 crashes in any game

Motherboard: GIGABYTE Z87X-UD3H

Processor: Intel core i5-4590

Graphics card: AMD MSI Radeon RX 580 8gb

Power Supply: EVGA 650 B3 80+ Bronze 650W

Overall problem:

My computer has frozen when under any load when playing games, will hard freeze the entire computer, with no warning. had some BSOD's in the beginning caused by un-updated drivers. before and after the BSOD's the computer would freeze entirely with no warning and no screen visual glitches 

 

my MSI Radeon RX 580 8GB crashes when playing any game, it has only crashed in games and never when using chrome or anything else, benchmarks continuously will not crash the system. games even in the start screen will crash within 5 minutes to an hour and rarely up to four consecutive hours. under heavy use in games such as Space Engineers, Beamng.Drive, Terratech, Scrap Mechanic and even Heavily modded Garry's mod will crash at high settings but light games such as Undertale, Roblox, and Reassembly will crash it at minimum settings too,

Things i have tried (in order of trying):

setting game settings lower 

playing lighter load games

running games separately from steam (steam was the only one crashing constantly on games, later turned out to be any game with further testing) 

running games separately on a USB drive along with steam

Reinstalling steam

Reinstalling drivers (19.4.3 and 19.5.2 [knocked windows out entirely and had to do a system restore]

changing the fan speed in MSI afterburner to be higher under load

changing the core clock to be lower in case the power supply cant handle it (which is unlikely)

running it in safe mode (got nowhere as windows wouldn't open any games due to lack of drivers being loaded)

stopping literally every process but required ones and steam in the task manager, this worked for about four hours and crashed again. 

deleting every driver and AMD program with REVOuninstaller and retrying in case of driver/program clash

0 Likes
11 Replies
supermangik
Adept I

Does your Wattman gets restored due to system failure? This could be possible be the issue and how to fix it here:
https://linustechtips.com/main/topic/1071941-solving-the-wattman-issue-with-amd-and-gpu-crashing/

It isn't uncommon to have the graphics subsystem hang with these cards but frequent BSOD and the issue you describe don't sound like the standard fare of issues. I would think you likely have a bad GPU or Power Supply I would talk to the support department from MSI about an RMA or whatever tests they can offer to pinpoint an issue.

You could try your card in another machine first if one is available. If you power supply is old it would be highly suspect. Most those EVGA's are good but I usually buy the Gold rated not bronze so can't speak to them. The gold IMHO are worth the extra cost. They save electricity, last longer and also make you components last longer IMHO.

There is a utility name OCCT from ocbase_com you can run to test these components too.

I would add that normally setting the power limit to 50 and using a custom fan and temp curve to kick the fans up high earlier usually fixes issues with these cards but again you issue sounds different.

There have been a few complaints though with these last couple drivers but people are saying the new 19.6.1 driver is better. You certainly could run DDU and try that new driver. I will share my power limit settings and fan curve with you from my RX 580 that works great. NO promises on it helping your situation but lets hope.

Edit: My thanks to another user that pointed out that I should have mentioned. You should not follow the memory or core clocks from this picture only the fan and temp curve and the power limit setting.

I would add that lowering the clock as suggested by another can certainly add to the stability of your card if raising the power ceiling alone does not help by iteself. However at that point you are clocking lower than defaults and not getting what you paid for in performance. So I would only suggest doing that if returning the card or an RMA is not an option. If not it is a good course of action and will help stability. I have found that setting static power states can be hit and miss on the stability from game to game. So remember all of these settings do not have to be made globally. You can tweak those settings in the individual game profiles too as needed.

Good Luck and sorry I was not more clear to begin with!

Thank you for the reply

The power supply is new and ill try the fixes you have mentioned

Do not do this, change the settings like person above.

If you're having BSOD, hangs, and restarting... It means it's either a driver issue or hardware issue. And often than not, it could be the problem with Overclocking. Having your memory clock higher in state 2 will only exacerbate the problem, not fix it. Underclocking your core will be more helpful to troubleshoot the problem.

I suggest you change all of your core to underclock it, and reduce the memory clock to state 1 level. Adjust your power limit to +50%. Adjusting the power limit to +50% will help the card to draw in the necessary power it needs to run properly, sometimes this is the case of the black screen and GPU freezing or stutter performance.

Underclock your all of your state GPU by lower than 100 mhz than your original core clock. The only state you can't clock is state 0. Just for some extra info, this seems to be an on-going problem with AMD. A failure in the driver or manufacture vBIOS that has not been addressed.

Since your card was introduced 2 years ago, I suggest you go to your card manufacture website (rather it be EVGA, Sapphire, etc...) and look for the updated vBIOS for your GPU. Some manufacture do not post this vBIOS update on their website and you'll have to create a support ticket to ask for it. If they refuse to give you the updated vBIOS, the next best bet is going to techpowerup website to look for it.

If you do plan on updating your GPU vBIOS, I suggest you download Techpowerup GPU-Z and save your original vBIOS to your HDD. Make sure it is the manufacture vBIOS release or you'll be screwing up things that you don't want too.

Download Driver Easy and make sure to update your chipset driver. I am not sure what the name is labeled for your chip driver, but it should be a file that either says chipset in it or Adrenaline in some cases.

0 Likes

Who are you talking to? I am having no issue with my card?

0 Likes

You're giving an advice to solve a problem by giving him your settings where you ramped up your memory clock state 2 to almost 3/4 with all core states ramp up higher than it's original core clock. Yea, I am talking to you when I said, "Don't listen to the person above." Your settings will only make the problem worst, not solve it.

0 Likes

I normally do add to only use the temp and fan curve and power limit settings. My apologies I did miss this on here. Some times I post late at nigh or at lunch at work when I probably am not proofing what I wrote well enough. I would add that MOST PEOPLE as you saying something proves know that this part is card specific. On the hundreds of times I have given this advice I think it is the first time I have omitted this. The OP is always welcome to follow any or none of the advice received here. We are not AMD professionals only other users.

I for instance would never under clock the core. If the power limit does not fix this then you have a faulty GPU and it should be RMA'd not cripple performance by lowering core count. I'm not saying it can't make an unstable card more stable but at that point the card is below spec and you are not getting what you paid for. I have found not leaving power states to be dynamic to be very unstable with Polaris. That doesn't mean another person doesn't have another experience. Most of us trying to help, only have experience with our own cards or worse share experience we have seen from others time and time again. Do to AMD not really helping in these forums in a large way, it is often the best help you get. 

Either way the best support for sure comes from the manufacturer or AMD themselves.

Remember we are all only trying to help one another. My intention is to never steer people wrong .

0 Likes
ludochat
Journeyman III

I think isn't uncommon to have the graphics subsystem hang with these cards but frequent BSOD and the issue you describe don't sound like the standard fare of issues. I would think you likely have a bad GPU or Power Supply I would talk to the support department from MSI about an RMA or whatever tests they can offer to pinpoint an issue.

0 Likes
jdrobinson314
Challenger

The jist of that external link was to switch to the normal vbios from the power performance bios.

Might be worth a shot and dial back any weird overclock MSI might have and build your own on the normal bios.

0 Likes
Hardwood
Paragon

Ok Ok, I'm late to the party but incase you haven't solved this yet..

1) I don't see that you've run a memory test, your symptoms sound like a memory error as the mem space fills to certain point.

2) Is there anything in the event viewer from just  before the crashes?

ASUS TUF X570 wifi, AMD Ryzen 5 5600G, ASRock 6700XT Challenger, Antec Truepower new 700w. 32GB
0 Likes
sacstradamus
Journeyman III

Same issue. MSI Radeon RX580. Gigabyte b450n ds3h. Ryzen 5 2600. First build. Any help would be dope

0 Likes