Hello, I've recently purchased a ASUS TUF 3 Radeon RX 5700XT and ever since putting this into my PC I experience game crashes / blue screens of death & just random shut downs. I have the following system, Link to videocard from where I purchased it, https://www.umart.com.au/Asus-TUF-3-Radeon-RX-5700-XT-EVO-Gaming-8G-OC-Graphics-Card_56055G.html Operating System - Windows 10 64bit CPU - AMD 1600X 6core 3.6ghz Motherboard - ASrock Fatal1ty X370 Gaming Series RAM - Corsair DDR4 8GB PSU - Thermaltake Smart Series 750W The card I was using before I purchased this one was Nvidia MSi 1060 6GB and I never experienced any of these issues. I've ran temp checks to ensure that the card isn't overheating which is certainly not the case, and I've never overclocked or modified any performance settings on the card, have always left it as stock as it came. Bluescreen of death reason - VIDEO_TML_ERROR - After googling it appears this message displays when there is a dysfunctional graphics card and/or driver. I've tried the latest drivers, I've tried rolling back to early drivers, I've tried using the recommended most "stable" drivers after googling which is 19.2.1, Does anyone know what would be causing this and/or a fix to the issue? Let me know if there's any other details I can provide, Thank you
Thank you for your replies, greatly appreciated Can I ask the best way to check my memory temps & hot spot temps? So far the very second my PC will crash / turn off I'll jump straight into BIOS upon restarting and check my temps and my GPU will sit around 70degrees celcius? I've still got 12 months return to base warranty (through Umart) and 3 years via Asus I believe with the video card, but before I go down that channel I thought I'd try and do as much testing / trying to fix it myself first if I can... don't want to be an annoying customer. Also, is there any recommended programs to install that I can use to stress test my GPU? I've done a quick google search and I'm going to try "FurMark" but I thought if I could possibly have something recording my GPU temps in the background while I stress test, then this would be a good way to get a gauge on if this is the issue? Thank you again!
GPU-Z is what I use, for stress-testing outside of known games particularly DirectX 12 or Vulkan API ones it's OCCT. https://www.techpowerup.com/download/techpowerup-gpu-z/ https://www.ocbase.com/ Asus and TUF I believe has had contact issues with some of the VRM's due to poor fitting or the thermal pads some users also do it on their own with the "washer mod" but tightening the cooler is something that should be done with care as to avoid PCB or components on the PCB getting damaged plus if it's a problem with thermal pads it could require replacing these or re-pasting the GPU or any number of other possible problems.
Driver wise I would clean it out and start with 19.7.5 and see how that performs then move to 19.12.1 (Not 19.12.2 / Adrenaline 2020) and see if that also works before going to 20.2.2 after that and if that's still stable 20.4.2 https://www.amd.com/en/support/kb/release-notes/rn-rad-win-19-7-5 https://www.amd.com/en/support/kb/release-notes/rn-rad-win-19-12-1 https://www.amd.com/en/support/kb/release-notes/rn-rad-win-20-2-2 https://www.amd.com/en/support/kb/release-notes/rn-rad-win-20-4-2 Otherwise it'd be the latest available and keeping it up to date from there. https://www.amd.com/en/support/kb/release-notes/rn-rad-win-20-8-1 Several but not all of the early driver stability issues should be resolved in the July / August drivers now but not everything and I have a feeling some of the lingering user problems can be traced back to hardware issues like the memory modules alongside some third party models having problematic construction and poor cooling characteristics which some manufacturers resolved and others just re-released their existing models and then when you buy the card you either have to see on the sticker on the GPU itself if it's a Rev 1.0 or Rev 1.1 or something or there might not be any indication at all and you just have to hope it's not some earlier model that has issues. My own card is a Sapphire variant of the Pulse type but Sapphire recently made a worse model of it called the Pulse BE or some such initially resembling a Black Edition but with several drawbacks including worse binned GPU chips so it's more like the Bad Edition so now users have to be careful of that too. EDIT: Oh! And for OCCT type testing any one single error is a problem it should always be zero. Hardware wise it's also useful to test without the DDR XMP profile active, card does a lot of stuff which stresses the RAM more so if this works you have to re-tune the memory profile. XMP's are a overclock and not a guarantee plus the on-die info sometimes just include primary and secondary timings with the motherboard memory training and compatibility having kit profiles and testing to fill in the extra info like the tertiary timings and more. Bios updates really help including the 300 and 400 series not just the newer processors and the 500 board models so if available and if it's stable the latest bios and AGESA code is heavily recommended.
Here Apoc_ https://www.techpowerup.com/download/techpowerup-gpu-z/ this can monitor almost all aspects and log them also for your trouble shooting attempts with your video card. But I must say that if the temp of your video card after a crash and a reboot into bios is 70 degrees your card has a severe cooling issue. By that amount of time even a reference cooled card would be down close to desktop idle temps.
@Apoc_ I don't have any personal experience with this model of GPU, but I have seen some videos that detail why it is overheating and what issues that will cause. 1. Worst 5700 XT Period, Asus TUF Gaming X3 RX 5700 XT OC, DON'T BUY! 2. Fixing the 5700 XT TUF, How Much Cooler Can We Make It?
Please check this, Asus as failed over and over with AMD cards. If this is the case you can RMA or fix it yourself.
My vote is for RMA. Get rid of that failure of a card. (Nothing wrong with 5700 XT chip itself, just this particular implementation from ASUS)
Rumors are that ASUS pulled this card from their line up and replaced it with an improved model . This one must have been old stock that the seller didn't go to the trouble to pull from inventory. I vote you see if you can rma this model for the improved model with ASUS. opps my bad this appears from the picture to be the improved version with the two different sized fans.
This card was very problematic when it launched. They fixed it on the hardware side, but it could be that you got an old one (even if the picture says it should be the evo version, doesnt hurt to check the card itself if it is indeed the newer version). Lots of people here say to check temps, which I agree with. If you see any temps going past or getting close to the max values when under load (or not under load), you know it is the card. And indeed, make sure you do not have any RAM oc or CPU oc when testing to make sure it is not an unstable overclock. I had a stable overclock for a year or so on my RAM and cpu, and had some problems with my new 5700XT. I never thought of the RAM or CPU oc because I almost forgot I had oc'ed everything because it had been stable for so long before I had the 5700XT. But I guess the 5700XT stresses RAM and CPU a bit differently, because indeed, when I had reset my BIOS to default/factory settings, everything was 100% stable, big issues and a lot of very small minor issues I barely noticed alike disappeared. Doesn't mean that you can never ever oc your RAM or CPU with a 5700XT, but you might need to redo your OC because the old oc isnt stable anymore in combination with the 5700XT. But even then, almost impossible to stresstest for every single scenario as chances are every game stresses your components a little bit differently. So if you encounter any instability again, always undo your oc's first to make sure it isnt that. Also, did you remove the drivers from your pc with Display Driver Uninstaller before installing the new card? It is not advised to use DDU every time you install new drivers, but this is exactly a scenario for which DDU is meant: a new card, and you switched from Nvidia to AMD, and you are heaving debilitating problems. So I would definitely try that (if the temps are fine. If the temps are bad/too high, than you can just stop there and you will probably have to RMA the card (assuming you are not running the card in a completely sealed box where no fresh air is getting in thus resulting in overheating)).
Also you might want to check into whether windows is having trouble with mapping your system memory with a video card with that amount of video memory 8GBs of vram as windows has only 8GBs system memory to work with. Windows use to reserve 4 GBs for its on use thus limiting the amount of memory needed for memory mapping for hardware use.
If you have the EVO version of this card, ASUS has revised the cooler design and it's actually one of the better 5700 XT's if you can get it at a good price. I've got the same card, I got it for $609 AUD so nice and cheap, and it's fine. The latest 20.8.1 driver was unstable in some games, but after rolling back to 20.7.2 everything is fine. If you have the older version, I would take it back, as it's not fit for purpose (hence why there's a new revised version).
