Meanwhile, I have now had two occurrences with the newest driver where the screen woke at the proper resolution and without having reordered my windows. I haven't kept an exact count but I'd say this is somewhere between 10 and 30 percent of cases (so far, for me) where it has worked properly. With the previous driver(s) it was a 100% consistent problem so it does seem they have tried to change something and may be on the right track of fixing it.
Latest drivers ar still a 100% failure rate for me. Been trying it for the last few days. As per, if you come back within 30 mins all is ok, if you leave it anything more than 30 mins the res issue is at 1200x800
Will try the next ones that come out
Did anyone have any joy with 21.4.1?
Been using them for a few hours but still got the screen issue on display port.
Just wondering how others got on.
It woke up as it should for me today at least once, however I am not sure if it was in deep sleep or not. Will leave it on overnight and report back in the morning (I turned it off last night as I got used to turning it off before I go to bed).
This morning my monitor woke up just fine (I did however had a GPU driver crash notification (was mining and playing with voltages and undervolting), but then I went for a walk with my dog and was away for about three hours and it woke up normally again.
Will test some more, but it seems AMD fixed the issue with 21.4.1 driver.
Too early to tell, but I haven't encountered the issue so far (n=4) on 21.4.1. Too early to call, though. It's been hit-and-miss on 21.3.2 and I haven't found a way to repro it 100% so I'll give it a few days and see how it goes.
After 2 days of testing, I can confirm the issue has been resolved: no more DisplayPort wake up issues with new 21.4.1. Kudos to AMD, however I would prefer having it mentioned in the driver release notes!
So I grabbed 21.4.1 and installed over the top of 21.3.1 and still had the issue. I recently had an issue with my PC so I did a full rebuild of windows and installed 21.4.1 and all is working.
Could it be something with WHQL drivers as looking back through the thread it seems we get reports of it working properly on WHQL drivers, the optional seem to cause issues.
Only been using for a few days so will see what happens down the line but so far so good.
Yeah I'm probably at n=10 now and I haven't encountered the issue anymore since 21.4.1. So I was inclined to believe they finally fixed it. Now that I read that it seems to have been fixed for you too, I'm more sure. A lot of people won't be coming back to this topic unless the problem appears again. I'll keep an eye out for it but indeed, looking good.
As for WHQL: those drivers aren't different from non-WHQL drivers. The WHQL process is just a certification, a sort of "guarantee" that those drivers are good and compatible with Windows, and it has nothing to do with the inner workings of your drivers. Getting your drivers certified means running a bunch of tests and sending the results to MS, who may take up to 30 or so days to check your results and afterwards you may release them as WHQL, and they may be included in Windows Update. Because this process takes quite a long time, driver developers often choose to release non-WHQL intermediate versions because they can be immediately available to the end-user when there are some hotfixes that need to be released immediately, or low-priority fixes that don't really warrant going through the whole process again while bigger fixes are still on the roadmap.
Remember, there have been plenty of drivers in the past months that are WHQL but still exhibit the problem, e.g. 21.3.1 which was your previous driver.
I'm glad it (so far) seems to be fixed for you, too. I'll keep an eye on this thread for a little bit to see if this is still a problem for others. In the meantime, other people who experienced this problem, please reply whether or not it was fixed for you with 21.4.1 and what screen(s) and graphics card you are using, and how they are connected. I'll try to compile the results and send them on, if necessary.
Vidko, as for not including it in the changelog, that should have been there yes, unless it was fixed unintentionally