PC Graphics

[belgiangenius]

Hi,

This summer I built a new PC:

Asus X570 TUF Gaming MB

AMD 3900X processor

32GB GSkill 3600Mhz RAM

Seagate Firecuda 2TB SSD

New Corsair RM850x PS

To start, I used my old 1070Ti video card.  It was rock solid in all games.  Even if you overclocked it to its limit, it would remain rock solid in all games.  So, all the other components in this build present no issues.  The system is stable.

Then, recently I decided to upgrade to a Gigabyte 5700XT OC card, and the weirdness begins.

At its default settings, you can run the 3DMARK Timespy stability test, over and over, and it will pass.  But in games, there are issues.  Forza 7 will crash or feeze at least once every two hours.  Project cars will occasionally crash or just give you a black screen when the race is supposed to begin.

Trying to overclock gets even funkier.  I can increase the power tuning +30%, undervolt to 1150mV, and again it'll hit ~2050Mhz consistently and loop the Timespy stability test forever with no problems.  Project cars doesn't seem affected by these settings too much (apart from the random weirdness it always exhibits with the 5700XT), but if I load up Forza 7 with any power increase at all (even +10%), it'll freeze almost instantly - even in menus.  Sometimes it'll work fine until I crash the car in the game, and then it freezes.  The weird thing is, Forza 7 is not a demanding game, and runs 4K 60Hz Ultra settings and only uses the GPU at 60% load - yet it's crash city.  Project cars 3 is very demanding but runs much more stable with overclock settings.

And Wattman seems to be another jumbo headache.  It just randomly keeps resetting my voltage back to the default of 1200mV.  What's up with this?

But even at default settings, this GPU doesn't seem entirely stable.    I'm almost ready to buy a new Nvidia card and chuck this thing.

Anyone got any suggestion?

[ajlueke]

It also does appear as if there is a quality difference between some of the different models.  The ASUS Strix and Powercolor Devil series don't seem to report as many issues as the reference based cards do.

[alseu]

Powercolor RX 5700 series graphics cards have the highest RMA rate according to the recent Mindfactory report. Specifically Red Devil models.

They share no info about ASUS cards. But there goes the infamous fiasco with a lack of GPU to cooler pressure causing overheat in Strix models. Also there was specifically bad cooler design for TUF RX 5700 XT leading to very high GDDR6 temperatures (over 100C at stock settings during load). 
Here is the screenshot taken from the Hardware Unboxed review:

My experience is mostly good with Navi. I use reference 5700 XT since launch. Haven't got any major issues anywhere but in one game. There were random black screens during Dark Souls 3 playtime. The game has the 60fps limit. At 1440p it results into a very light to average load with this GPU. Random blackscreens could occur only at the time when I wasn't streaming or recording. That seemed to be a hardware issue with lower C states but I couldn't reproduce it anywhere else. Finished the game with enabled instant replay and decided to leave everything as it is for now.

[ajlueke]

Good info to have!  I am new to Navi, and have been using a Powercolor unit.  No issues to report so far.  

https://www.userbenchmark.com/UserRun/33706830 

[koszalek]

AMD driver wersion?

[ajlueke]

20.4.2.  I haven't moved to the September driver yet.

[pokester]

Based on these forum the latest driver seems to be fixing a lot for users so I might be worth moving to that driver soon. Sounding like RDNA2 may actually have a pretty good driver to start out with. 

[belgiangenius]

The only problem I still have is that Wattman still keeps resetting my voltage to 1200 mV.  All other settings in Wattman stick, except that voltage that keeps getting reset to default.

WHAT IS THE DEAL WITH THIS?

[belgiangenius]

So, after using the latest driver version for some time, I must say that stability went from COMPLETELY UNSTABLE with the previous two versions to about 98% stable now.  I still get the occasional crash in Project Cars 3.

But Wattman seems like complete cow dung.  It randomly changes my voltage settings.  Sometimes you make a change and apply it, and it completely ignores the changes and leaves it as is.  What is up with this Wattman POS?

[pokester]

I can't use the current interface it is IMHO garbage. In fact on my RX580 I can't even get the 2020 drivers to load. I have several cards that are still supported but can't use the current drivers unless you load the driver only from device manager as the AMD installer does not allow this. Then you can either still use the 2019 interface if you had it install or just clean install the AMD driver only through device manager and control it with 3rd party softare such as Afterburner which seems to work much better. Just don't I stall the Adrenaline software and Afterburner as that just adds to the instability nightmare. 

[belgiangenius]

Thanks for the suggestion.  I tried that and am now just running bare 20.9.2 drivers with MSI afterburner.

Unfortunately, it did not make much difference.  The last version of the driver was a HUGE improvement in stability, but most games still crash every 1-3 hours or so.  Oddly, games rarely crash during the game.  It's usually in the menus or during some transition.  Sometimes the game will crash and eventually shut down, sometimes the system freezes, and sometimes the system reboots.

[pokester]

Sorry that sucks. I have recently seen in a couple tech videos that Navi was released in fact with hardware issues. While the AMD employee was not named that confirmed this the interviews were with an anonymous AMD employee. 

I have thought this from the start but and it sure makes sense as to why there is such extremely large amount of issues with these Navi cards.

On the good news side this AMD person did say that the problem is fixed in RDNA2. Let's hope so. 

[satanael]

pokester what you said doesn't make any sense i still believe it is driver issues since my GPU under heavyload can reach 144hz easily without Crashes on the other hand my GPU can't Work at 144hz With Some Desktop apps.

[pokester]

Not sure why it doesn't make sense but okay. I only repeated what I read on a couple tech sites as I already said. No my words theirs. However it makes perfect sense to me as to why so many issues still are not fixed, as drivers can only do so much in some cases to mitigate hardware bugs. Hence why well known flaws like spectre and meltdown for instance are so debilitating. I just found it interesting sorry if it makes no sense to you. 

[pokester]

Also I should add that I never said it was all hardware issues just that hardware issues also exist that may or may not be fixed by software updates. I litter-ally made all the news outlets just a couple days ago that Navi has security flaw that will not be fixed until quarter 1, 2021. I think it is very possible this is the hardware bug the anonymous source was talking about a couple weeks back that I saw on the tech sites. Don't know for sure but it seems it could be as AMD admits the issue causes BSOD. 

[belgiangenius]

Given my experience with the card, I'm not surprised.  I've never owned a graphics card that crashed so much in games, even at stock settings.

I have a few computers in the house, so I'm always upgrading something.  I'm leaning to buying an Nvidia 3000 series card, assuming their fake launch ever makes the transition to being a real launch and product is actually available.

I'm skeptical of AMD based cards now.  If they would make a fair offer, like anyone who has stability concerns with a 5700 based card can turn it in for a 50% discount on Big Navi, I'd probably take the chance on Big Navi, but otherwise I think my future will lead back to Nvidia.

[satanael]

belgiangenius try this

144hz Refresh Rate Cause Hard Crashes Over Displayport? (Workaround) 

[colesdav]

Try this: 
https://community.amd.com/t5/drivers-software/amd-driver-fix-and-install-using-ddu/m-p/426238

I have been running RX5700XT and they seemed stable for August ~November drivers. 
Almost 3 months!

But the RX5700XT I currently have installed has been crashing and Blackscreening on recent drivers up to and including 20.11.3.

I have just moved to 20.12.1 drivers. 
I am running Windows 10 Pro 2004 with latest patches. 

It seems like Windows 20H2 may be causing problems for AMD Drivers based on reports on this forum. 

Check how much the AMD Drivers are boosting the GPU Clk versus the specification for your GPU on Manufacturers website. 

Try increasing the fan speed and the Power Limit from default. 
 
Good luck. 

[Paaaaaja]

I have RMA my Aorus RX5700XT 3 days ago. Just after I DDU 20.9.1 and installed 20.11.1 in mid of November to get rid of random in-game stutter, I got black screen crashes - especially when I lowered power limit (then it was within minutes). I have tried almost all drivers from 20.5 through 20.7, all since 20.9.1 till 20.11.3. 2x fresh Windows re-installs, many DDUs, nothing helped. So I RMA. I hope for positive outcome:-/ 

I am running Powercolor Red Dragon RX 5700 XT now, was able to snap one in the current GK crisis - now I am 4th day without a crash - hope it will stay that way.

I believe most of the issues described here are hardware ones. I am still curious why my card was running OK till DDU, but this is now beyond my reach.

[colesdav]

They are overclocked out of the box to within an inch of their life to try to compete with Nvidia. 
Usual story with AMD GPUs. 
Some AIB models have Ram running too hot as well, because of poor thermal pads / heatsink design/ coolers falling off. 
Drivers were unstable with blackscreening from launch until August 2020. 
Was o.k. until 2020 20.11.1/2/3. 
2020 20.12.1 seems ok again. 

[idiocracy]

I have an RM750, two individual power rails to the card. But i have similar issues. I didn't consider the windows update to be the cause.

But basically i lowered clocks to 1900mhz, but regardless, it still crashes at 1175mV. I just can't seem to get it stable. But if windows is the issue here that would explain a lot.

[Charred]

I had the same problem for awhile, I tried everything. used DDU multiple times and tried all of the previous drivers, I even only installed just the drivers but nothing worked and even underclocked it to 1999 frequency with 1100 voltage and still nothing worked. Finally i saw a post about underclocking the gpu to 1900 frequency with 1000 voltage its stopped all of my crashes and my temprature arent high. though every time i tweak the other settings it would make my games crash so i wouldnt recommend tweaking the other settings. I did tweak the fan speeds a bit to get lower temps 

[alyzee]

The suggestion you write to fix GPU in AMD standards causes most users to have more problems probably. Every GPU has an operating voltage range. GPU crashes instantly when voltage is applied outside the operating range. So overvoltage won't hurt a GPU. However, excessive voltage may cause the GPU and VRM to overheat.

[alyzee]

I was also using a Corsair (a 750 Gold) a long time ago. It was used with FX9370 and HD7870 (approx. 450w) with a system suitable for its capacity and it was burned. Those who buy Corsair think they are getting a quality product and this has nothing to do with reality. For this reason, I just warned you.

[josephweasley]

I am a user of Sapphire Nitro + RX 5700 XT for a year and a half that i have been trying to solve the problem of black screens in the game, now always with this error (We detected a driver timeout has occurred on your system). I have opened various posts on this forum i have also tried to undervolt but with absolutely no luck. 

In the last post i opened (https://community.amd.com/t5/graphics/we-detected-a-driver-timeout-has-occurred-on-your-system-rx-57...) this user claims to have crashed only once with this error. As i also read about this thread could the power supply be my problem? I have a Corsair TX650M 80 plus bronze, for example this user has a 750W 80 plus platinum, as @teoretic wrote a better power supply (80 plus gold or platinum) could help solve this problem? Or maybe or me a bad card? Please help!

Undervolt settings: