Ahh, I was asking out of curiosity because there is a known issue with my specific model of Asus RX 470 that is basically what you described and was wondering if the 480 had the same issue as well but I guess it's not the case because it's MSI. For my card it's that Asus have set the state 7 voltage too high, I have to actually undervolt the card so it doesn't lose signal with the monitor when playing video games. Perhaps something for you to try? just the state 7 voltage, try lowering it about 20mV ? Just in case you don't know where to look. Spoiler Not sure if this will help, but hey, something to try at least.
i've told you it's not card. i have same problem on 2 completly different rigs. it's 100000000000% driver issue, it was reported on AMD forum years ago a no1 managed to indetify reason (or hasn't try). all you can do is clear drivers with DDU and instal new recommended one every time they appear, with fingers crossed it will happen less often (no it will not end). if it appear less often, stick to this driver and do not touch it until you must. and again this problem does not appear in Linux. that single fact is making all hardware theories invalid.
i'm million % sure. i had this issue on my r9 390 with hdmi cable and 19" monitor, now i have vega56 with dp cable and brand new 27" monitor and same problem exists. no solution, no idea for solution. windows driver issue, no more, no less. you just can have hope that most recent driver will have it less often.
So the same problem occurred with different cards? My guess would be PSU not able to provide the current fast enough. Have you been using the same hardware between GPU upgrades?
my guess is, neither of you bother to read all my previous posts. nothing better than "advices" based on last post. 2 different card, 2 different monitors, 2 different computers.
Woah.... calm down, are you using any shared hardware between build? It can't be a coincidence that you're having same issues with different GPUs.
I'm calm, just what's the point to repeat same thing over again? No shared hardware as said before as well. 2 different rigs. And again this problem doesn't exist in Linux. My instalation is using mesa, not AMD driver. That's clear indication its driver related.
Only trying to help you here.... I'm asking for clarification, you didn't state what was different in the other computers you used your cards in, whether or not "new rigs" had parts from previous builds (such as using the same PSU but everything else different, or using the same hard drives etc). I understand it's frustrating having a problem that you can't solve but please show a little patience. Does this also include peripherals that you're using? We need to rule out everything we can. (I had a USB hub that would crash my computer if I plugged a certain USB stick in, it was very frustrating to think such a thing could happen). Now yes it's clear drivers are showing a problem however, going from my experience for example, older drivers with my R9 Fury, the card used to draw more power but as newer drivers came out, efficiency improved. I'm still thinking this is likely a power delivery issue. My next question is, what are you using as your electrical outlet? Do you have a surge protector? Are you using a multi-plug extension? There could be noise in your electrical supply. One reason why the problem is maybe exhibited in Windows vs Linux is because of the difference in efficiency in the drivers, windows driver may be having the card draw more power or may have more aggressive power changes etc. This could also be the infamous "black screen bug" which has been present throughout various iterations of GCN hardware and their derivatives - it's usually related to memory clocks changing which sometimes bugs out and cuts feed to the monitor. One thing you can attempt is to lock memory to it's maximum clock, there has also been problems with Wattman not applying profiles saved each startup. Although the new drivers are using a new method for overclocking, Wattman profiles can still cause issues with the new software, particularly if there are any registry keys that haven't been removed (I wouldn't be surprised if the overclocking tool is still causing issue like wattman did either). So, drivers could indeed be triggering an issue but there ultimately could still be and underlying problem.
2 completely different systems, every single part changed. Few years in time between 2 builds. 2 different houses. Older system had this issue way before wattman was introduced and was never overclocked. New system with vega have this issue with default settings as well as my custom under volt plus manually set power limit. Problem appear rate is random with every new driver. Some driver has it few times a day, some few times per hour. Driver installation always done after clearing with ddu, as well as new windows installation. Nearly impossible to identify reason. I've tryed probably every suggestion I could find online over past few years. I was one of two people reporting such issue with older AMD cards. Only asnwer AMD ever gave was that they were not able to replicate problem to try and solve it. This post is first i see in past few years with exactly same issue. Most interesting is, that it takes place when card is not under full load. However playing same game under Linux does not give any issue.
I'm thinking it is the black screen issue then and you're unfortunate to encounter it several times across generations - as far as I'm aware, it's occurring when the memory switches from low power saving clocks to performance clocks. Vega saw it occur more frequently and it's happening with Navi too, any card that has memory that switches clocks. And yeah this is an ancient problem that's been dogging GCN for years but it's become more prevalent with newer cards that push for more aggressive power savings. Try locking memory clocks to their max defaults and see how you get on? (So change idle memory speeds to the same as performance speeds, just the memory, not the core clock). Edit: Thinking about this though, I wonder if the black screen issue is a compound problem? Perhaps it's not just the drivers at fault, perhaps it's Windows too? Maybe Microsoft need to step in? (This question/speculation is aimed towards the general black screen issue overall)
There is a known issue "known in this case referring to Asus acknowledging the issue and offering replacements for affected cards" for the ASUS ROG Radeon RX 470 Strix Gaming OC 4GB, I decided not to replace the card and instead undervolt state 7 voltage which fixed the issue, the issue in this case is not related to the driver, it is also not related to the more current black screen issue that is present in the latest few drivers released, which I am also having if I use those newer drivers, I am currently stuck on 19.12.3 until AMD sort out this separate black screen issue. Because op has a 480 I was wondering if maybe that model had a similar issue to the 470 but since it was not Asus branded I am not 100% sure, since he has not responded I am wondering if he has tried what I suggested and if it made any difference.
The guy stated on the previous page that things are fine with the dp cable. What issue are you having with understanding that? How was that wasted money? Am I missing something?
my r9 390 is from Asus, however my Vega is from Sapphire. i did undervolt my Vega, no change. tha same guy said on previous page that problem is not gone, and after test with 2 monitors and different cables problem still exists, only scenarion changes slightly. please read before you post. "What issue are you having with understanding that?"
If you are having the black screen being caused by the new few driver releases try driver version 19.12.3 or 19.12.1
again, this problem exist for me for very long time. changing driver version doesn't solve it. just randomly it's get reduced or it intesify depending on driver, no idea what is the reason tho.
