Graphics Cards

[mmikko_]

Recently built a new computer and decided with an RX 5700XT as RTX 2070 Super was more difficult to get for me and the extra $100 CAD wasn't worth ray tracing or the near identical performance in the games I'd be playing at 1080p.

-Basically I have had lots of trouble getting this GPU to work properly without crashing my computer or recently its been only freezing and I feel there's nothing more I can do. Also I have only had this computer for little over 3 weeks on a clean windows install.

My specs:

CPU: R5 3600(non X)

Mobo: Asus X570-P

RAM: G Skill 2x 8GB 3600MHz 16CL (Used DOCP in BIOS to achieve 3600MHz)

GPU: Gigabyte Radeon RX 5700 XT 8 GB AORUS Video Card

SSD: 1TB Sabrent Rocket NVMe 

PSU: Corsair CV 650 80+ Bronze

Monitors: DisPort Acer XF240H(Primary) & HDMI Acer P206HL(Secondary)

*All temps are Celsius*

-I've tried numerous "solutions" that have seemed to work for other people or just things with my knowledge that are red flags to crashing or caused errors in Event Viewer that I knew how to deal with myself.

-I have in the past few days of my troubleshooting and "fixing" caused my computer to stop full on crashing during a game but now it simply freezes it will still play any audio like that of the game or spotify/youtube but the screen will keep displaying the screen it froze on for at least 5 minutes as I've not waited any longer before restarting.

-My temps seem okay, GPU temp on Adrenalin and HWMonitor Idle at 49(lowest I've seen)-56(highest I've seen) and during gaming anywhere from 64-high 70s maybe low 80's, haven't seen anything really above 80 lately. Junction/Hotspot temps are about the same as GPU idle but reach 85-110(not 110 as often since new case fans) during load

-CPU Temps haven't reached anything higher than low 80's since new fans as well

-Wattage during freezing is normally quite low(80-160W) not often anywhere near the 225W the GPU is said to max at and 650W should be more than enough

Things I've done/tried: 

- DUU'd drivers, fresh install (Not turning on gaming mode said casual or custom)(Multiple times)

- BIOS Updated to latest

- Chipset Drivers updated(2x)

- Fixed all things causing errors in Event Viewer (ATI External Events Utility, Headset driver that caused errors)

- Windows Update

- Set PCI to Gen 3

- Set High performance modes

- Re-seating GPU
- No Hardware Acceleration (Discord, Google, & Spotify)

- Checking drives for corruption

- Graphics settings in Adrenalin

      - Anti-lag OFF

      - Chill OFF

      - Boost OFF

      - Img. Sharpening OFF

      - Enhanced Sync OFF

- Freesync OFF

Games I've played so far(All 1080p):

- Divinity Original Sin 2: Definitive Edition (Max settings) (Freezes very frequently)

- Rainbow 6 (Low sett.) (Freezes as well)

- The Witcher 3(Max) (Hav en't spent an hour in one run so far but no freezing)

- Monster Hunter World (Max) (Been a while since played(before troubleshooting) it was the first game to crash)

- Minecraft w/ Shaders (No freezing(Little GPU usage))

There's probably more but I will be open to suggestion if there's more or might've been something I overlooked. Regardless if none of this works I'm posting this in hopes that not only I find a solution for myself but things that I've tried may be the answer for others.

I am iffy with messing with tuning, under volting and fan curves as I want to keep my warranty intact as a last resort if I can't fix it but I very much don't want to have to resort to a warranty.

 *At some points Event Viewer would show "Fatal hardware error has occurred"

[mstfbsrn980]

The source of the problem is unclear. You can try the following things to find the source of the problem.
+ Lock the CPU speed to 4Ghz by giving the manual 40 multiplier to the CPU Core Ratio with the BIOS. By following this way, you can test for the CPU.

+ Reduce the RAM speed to 3000Mhz with the BIOS. By following this way, you can test for the RAMs.

If I was in your place, I would try the two ways above separately. If improvement occurs, mobo or RAM is faulty. I think GPU or PSU problems will not cause freezing in general. If you are very sure that your CPU and RAM are stable, I can write trial methods for GPU-PSU testing.

[mmikko_]

Would locking the GPU speed or anything conflict with any warranties?

I just want to be very careful with them as I'd rather not spend anymore money than I already have.

Do you think it could possibly just be AMD's drivers acting up?

Thank you for your help I will try those as I haven't touched those kind of settings

[mstfbsrn980]

I have never provided you with advice that will affect the warranty of very new products you have purchased. Secondly, I recommend that you apply the first things I wrote because you did not share any information that would be a problem with the GPU. As the third. The settings are available for the GPU, but we need to find the problem. The problem is not obvious.

[mmikko_]

Unfortunately neither change seemed to fix anything or reduce the freezing, also notable I have run a few stress tests(FurMark, Prime95 and PassMark BurnIn about 30 minutes to an hour each) and they all ran smoothly not showing any signs of error or freezing. I felt that that it may possibly point toward a definite cause. 

Also in event viewer this warning seems to be there the time it freezes: 

The application-specific permission settings do not grant Local Activation permission for the COM Server application with CLSID  {2593F8B9-4EAF-457C-B68A-50F6B8EA6B54}
and APPID  {15C20B67-12E7-4BB6-92BB-7AFF07997402}
to the user **** SID (S-1-5-21-2089789035-3357418236-46045586-1001) from address LocalHost (Using LRPC) running in the application container Unavailable SID (Unavailable). This security permission can be modified using the Component Services administrative tool.

And another error that seems to show typically a few minutes or so before a freeze is this one:

The system has rebooted without cleanly shutting down first. This error could be caused if the system stopped responding, crashed, or lost power unexpectedly.

I'm simply trying providing as much information possible in hopes it may help pin point something

[mstfbsrn980]

No. It is very difficult to determine the source of the problem with the information you provided. You can run a similar improvement test for the GPU if you wish. However, you need to share a screenshot of global WattMan settings. WattMan settings should be adjusted manually and all numerical values should be visible for a screenshot.

[mstfbsrn980]

Edit: I also recommend using a single monitor for a while. Here are users that use two monitors and get similar errors.

[mmikko_]

So I hace had a spare GPU lying around for the longest time so I decided to try and use it and see if it changed anything. It's an MSi GTX 750 Ti, I am still using the dual monitor set just with some slight cable modifications. I've been playing Divinity(One of the games that freezes often for me) and its been 2+ hours without a single freeze. This has lead me to believe it is likely a GPU issue and I'm going to look into a warranty claim for a defect but before I take any action on it your input would be greatly appreciated and whether or not you believe it is the GPU.

[mstfbsrn980]

There is 150W TDP difference between 5700 XT and 750 Ti. So the problem may be with the PSU may be not. Everything can cause freeze with a PC. In general, a motherboard and a RAM cause this problem. However, even a broken USB device can cause system freezing.

[mmikko_]

That makes sense, I'm going to look into getting a return for my 5700xt and if my problem persists I will look into the PSU next if not I may use my motherboard warranty as I purchased an extended one, I'll just look into it more. Anyway thank you very much for your support and help it means a lot.

[steve8120__]

I am having almost identical issue.

[gigeog]

Same problem here, frame drops / memory leak. 

It is my first AMD GPU, and probably the last.

[mmikko_ · Accepted solution]

***UPDATE*** POSSIBLE FIX:

I am currently on the fence with following through with an RMA as I recently changed some things that seemingly stopped my issue. I have only used it after "fixing" it for a day so it may not actually work (hence why I'm still on the fence for the next few days). Basically one morning it black screened and crashed on me twice in Frostpunk. So I said screw it and went to Adrenalin and turned on all the things people say crash systems. 

What I did:

Before anything else I pressed Factory reset in the Adrenalin system settings and set standard preset, then

I Turned on Anti-Lag

Turned on Image Sharpening (100%)

Turned on FreeSync

And lastly turned on Enhanced Sync

This may work for others but for me it stopped the game from crashing my computer as well as others even though it crashed twice less than 10 minutes after playing. I still need to test and find out if this works and wasn't simply a fluke for the rest of the day, but it seems promising.

[notflawless]

Believe me, there is currently no final solution. Only AMD can fix the Problems. I spend two months of my life in this card, i tested all suggestions. Every time i thought it was fixed, the sporadic blackscreen say "Hello".

Today my card goes back to the seller. My Seller said: At the moment, dont buy an AMD Card, we only have problems with it.

I see dark clouds when i look to the RDNA2 release....

[trinimac]

For me my 5700 with the 5700 xt bios was crashing randomly daily during games. I have sent daily crash report for weeks...

I have a Asus TUF x470 gaming motherboard and that might be the cause of the issue based on the comments many have made here and there.

I have updated all my chipset drivers (which didn't help) and opened radeon software to see if I can tinker with my settings. I noticed that Auto Tuning was set to Off under Performance > Tuning so I set it to default. Since than I haven't had a single crash.

My settings on the tuning page (Performance > Tuning ) are:

> Tuning Control > Automatic
> Auto Tuning > Default

It might help some 5700 & xt users.