ye I also RMAed & went for the Rtx 2080 instead. all issues solved and works like a charm..
working perfectly with my ryzen 2700X though.
Thanks for sharing your experience.
Adrenalin 19.9.2 on brand new XFX RX Vega 64 Liquid and PowerColor Red Dragon Vega 56 have improved stability lots for me since I purchased the Vega 64 Liquid in November 2018 last year. I just purchased a second XFX RX Vega 64 Liquid. I am using the GPUs for programming and Blender. I also run Ryzen 2700X and Ryzen 3600 machines. The reason I have gone for AMD has been low cost of the hardware for the performance. Although the Vega 64/56 GPU drivers have been much more stable recently, I can understand your point completely. The XFX Vega 64 Liquid was very unstable for a long time. I initially thought it was a faulty GPU and almost RMA'ed it but it was definitely the drivers. I purchased an RTX2080 in December 2018 because I could not put up with the Vega 64 stability issues. Despite the RTX 2080 recently launched, it was and still is more stable than the XFX RX Vega 64 Liquid. No BSODs or PC Freezes or restrs from the RTX2080 yet.
The Ryzen 2700X build I did cost me lots of lost time and production because of memory compatibility problems, although that has definitely improved with Ryzen 3600.
I am very interested in RX5700XT but I am not going near them until ~ February 2020 because I think that is how long it will take for the AIB GPU Drivers and BIOS to be stable. The prices will hopefully drop towards ~ RX590 level by then, because that is where a GOU with this die size should be. It is really an RX680 after all.
Ryzen is rock solid, in fact my intel systems have had way more failures than my AMD systems ever had. I have however gone through two high-end boards for various BIOS related failures, the later being a faulty BIOS chip itself as opposed to unstable BIOS code.
90% of the complaints against ryzen's stability are user-error, things such as not loading XMP settings properly or not being aware of some necessary voltage bumps in some cases. The remaining 10% are BIOS bugs from shitty companies like MSI and gigabyte. Both of these issues apply equally to intel systems as well.
As for navi, not sure, what was the specific issues that you were having and did you run a DDU? (with ethernet/wifi disconnected to ensure winupdate doesn't break it) My VII has been pretty much rock solid, same with my 290X prior, HDMI driver 10.0.1.12 and .11 however do have audio issues that need a DDU+clean install to fix, presumably some registry settings that cant be reset by the traditional driver updating.
Similar to intel, I have in fact had more issues with nvidia in production and development than I ever had with AMD cards, particularly with their lackluster data type and format support in direct comparison. Funnily enough, one of the biggest issues I've had with nvidia is broken audio, and their driver's uncanny ability to break other audio devices in the process.
I was experiencing the exact same symptoms, intermittent crash, had multiple monitors and one was even a Dell using DVI. After trying all the settings I gave up and RMA'd my card convinced it was the Vega 64 design and ASUS sent me another Vega 64 back... Well i was pissed because sure enough it started crashing again! But after I took out the Vega 64 and put back in my old video card I noticed one of my monitors on display port was getting some weird distortions. So i put the Vega back in, same distortions... By trial and error I discovered the display port cable, or perhaps the port, was the culprit. I bought a new mini dp to dp cable for it, because it could take both, and I haven't had a crash for just under a month. So if you're having this issue and you are considering RMA'ing i would try checking the cables and ports too. Hope this helps someone.
Hi there, I've had the same issue for over a year I think, have been trying all the methods found in this thread over this time with some success but as others have mentioned, as soon as WattMan 'forgets' my settings I get crashes. Usually after an hour or so it will forget and crash.
But I think I've stumbled across a solution;
At least it's working on my system - 4 days with no crashes whereas it would not go more than an hour or two at a time before without black screen and driver crash before this.
(My card had previously been back to the supplier who tested it and found no faults so the errors were software, as we know)
Last Friday night I wiped all the AMD drivers for the card and let windows install as standard VGA adapter. Didn't do much good, the card didn't crash but I only saw 1 screen and no Rift...
So then I did a clean install of the very latest optional driver, 19.11.1 but I did not tick the AMD settings software in the installer. Just the driver and the HDMI audio driver. This way the AMD software and WattMan etc isn't even installed.
And voila. So far so good. Card works as it should.
I use afterburner to keep an eye on GPU speeds and temps instead and it hasn't crashed since.
(Edit - after writing this I had one crash last night. Will test again tonight.)
Do you have a freesync display? and was that crash last night with the rift running? it's entirely possible that not installing the settings UI means it's locking to a higher power state, and/or freesync etc is staying disabled, either of which can help (or worsen) stability.
But really, anyone still seeing these issues should just get the card RMA'ed and fully refunded, this is possible in AU, EU and UK even out of warranty (albeit the procedure is rather harder post-warranty). Not sure about the US, but you could always form a class-action? hell, they did it with the FX chips and won by default...
I can definitely confirm the hardware issue is non-existent on the VII as well (with the exception of a driver-specific issue I got once with a multi-cable display), the difference the dedicated SoC power rail on the VII makes, is as clear as night and day. I'm pretty sure first-gen NAVI has the same dedicated SoC rail as well, for the same reason (and despite remaining mostly unused as per the architecture block diagram).
Hello mate, thx for your info. No, I don't currently have a freesync display, I've got 1 ASUS in the middle and 2 HansG either side. all 24 inch, 1080 screens. All on a display port each.
I've been lookin into getting a bigger centre screen with Freesync. Would go for it if I could get the card to behave!
The Rift is currently unplugged in case of it being a cause or factor in the crashes; plus it would start itself up everytime it was knocked or something (cat) went near it cos of the proximity sensor haha and it would then try to take over everything I'd be doing and start up VR home and so on. So I keep its HMDI cable unplugged from the GPU when not in use.
As for the UI, I can see the card is still downclocking and upclocking in my afterburner, so it must be coded into the driver, I presume.
Last night I installed the latest 19.12 optional, again without the settings UI. It didn't crash last night or this morning. I thought to myself that it didn't start to crash (the other day) until after a few days/reboots since the driver install, could it be related? So I've not shut it down and just set it to sleep when I'm not on it. See if that might help.
RMA; I sent it back to Scan to RMA it, about 6 months ago. They tested it (not very thoroughly - just set it with resident evil playing I think he said and left it for 48 hours) and said the card is perfectly fine so it must be a driver issue you've got and sent it back to me. I told them it doesn't crash when being used hard, only when not doing 3D. But they sent it back to me. What would you do next? I could email them again and see if they would consider escalating the RMA straight to Sapphire?
Edit: Found the email from Scan, quotes "GPU tested over the weekend more than 48 hrs, Stressed under high load on assassin's creed odyssey @ 4K ultra settings - GPU never crashed during the testing process - no freezes. NFF"