Quoting: ShmerlIf anyone in EU is looking for Navi cards on sale, see:
https://www.mindfactory.de/product_info.php/8GB-Sapphire-Radeon-RX-5700-XT-Pulse-8G-GDDR6-HDMI-TRIPLE-DP-OC-W--BP--_1324347.html

I bought exactly that card from exactly that shop and I'm more than happy with it.
No issues so far. Arch, Kernel 5.4.

As you said there are important Navi fixes in 5.5, what improvements are (apart from overclocking support) to expected from the 5.5 kernel release?

The price for the XT version might have come down a bit lately. But for the ordinary 5700 version it has just increased noticeable again, probably because it was too close to that of the just released 5600 cards.

Quoting: subAs you said there are important Navi fixes in 5.5, what improvements are (apart from overclocking support) to expected from the 5.5 kernel release?

A few freeze fixes, but they possibly were backported to 5.4.x too, I didn't follow that.

This is a bad review of the current situation with the drivers on windows.
There some AMD fanboys that are saying this isn't so, but are overwhelmed.

Good opportunity for Windows AMD users to switch to Linux.

More than 3 months have passed now, so it's safe to assume that the situation as become better, right? Unfortunately this has not happened.
Setting up a Ubuntu 20.04 (yes I know - it's still two more months till release, but the symptoms are the same I had with 19.10).
No additional PPA, the stock Navi BIOS that comes with the distro, no additional "sensors package" installed, Kernel 5.4, Mesa 19.3.3. Just OOTB.

2 out of 3 boots give me

...
[    3.356136] amdgpu: [powerplay] smu driver if version = 0x00000033, smu fw if version = 0x00000035, smu fw version = 0x002a2f84 (42.47.132)
[    3.356137] amdgpu: [powerplay] SMU driver if version not matched
[    3.370207] amdgpu: [powerplay] SMU is initialized successfully!
...
[    6.885145] amdgpu: [powerplay] failed send message: PowerDownVcn (44) response 0xffffffc2
[    6.885185] [drm:amdgpu_dpm_enable_uvd [amdgpu]] *ERROR* [SW SMU]: dpm enable uvd failed, state = false, ret = -62.
...
[   11.748578] amdgpu: [powerplay] failed send message: GetMaxDpmFreq (31)  param: 0x00000000 response 0xffffffc2
[   14.224396] amdgpu: [powerplay] failed send message: GetMaxDpmFreq (31)  param: 0x00000000 response 0xffffffc2
[   16.705128] amdgpu: [powerplay] failed send message: GetMaxDpmFreq (31)  param: 0x00020000 response 0xffffffc2
[   19.185534] amdgpu: [powerplay] failed send message: GetMaxDpmFreq (31)  param: 0x00020000 response 0xffffffc2
[   21.617719] amdgpu: [powerplay] failed send message: GetMaxDpmFreq (31)  param: 0x00000000 response 0xffffffc2
[   23.985283] amdgpu: [powerplay] failed send message: GetMaxDpmFreq (31)  param: 0x00000000 response 0xffffffc2
[   26.462730] amdgpu: [powerplay] failed send message: GetMaxDpmFreq (31)  param: 0x00020000 response 0xffffffc2
[   28.940680] amdgpu: [powerplay] failed send message: GetMaxDpmFreq (31)  param: 0x00020000 response 0xffffffc2
[   31.203724] amdgpu: [powerplay] failed send message: NumOfDisplays (64)  param: 0x00000002 response 0xffffffc2
[   33.417301] amdgpu: [powerplay] failed send message: NumOfDisplays (64)  param: 0x00000002 response 0xffffffc2
[   35.906984] amdgpu: [powerplay] failed send message: NumOfDisplays (64)  param: 0x00000002 response 0xffffffc2
[   38.386750] amdgpu: [powerplay] failed send message: NumOfDisplays (64)  param: 0x00000002 response 0xffffffc2
[   40.866718] amdgpu: [powerplay] failed send message: SetMinDeepSleepDcefclk (34)  param: 0x00000010 response 0xffffffc2
[   43.346588] amdgpu: [powerplay] failed send message: SetMinDeepSleepDcefclk (34)  param: 0x00000010 response 0xffffffc2
[   43.346590] amdgpu: [powerplay] SMU11 attempt to set divider for DCEFCLK Failed!
...
[   47.525853] amdgpu: [powerplay] failed send message: GetMaxDpmFreq (31)  param: 0x00000000 response 0xffffffc2
[   50.000548] amdgpu: [powerplay] failed send message: GetMaxDpmFreq (31)  param: 0x00020000 response 0xffffffc2
[   52.476361] amdgpu: [powerplay] failed send message: GetMaxDpmFreq (31)  param: 0x00020000 response 0xffffffc2
[   54.689797] amdgpu: [powerplay] failed send message: NumOfDisplays (64)  param: 0x00000002 response 0xffffffc2
[   56.903180] amdgpu: [powerplay] failed send message: NumOfDisplays (64)  param: 0x00000002 response 0xffffffc2
[   59.379406] amdgpu: [powerplay] failed send message: GetMaxDpmFreq (31)  param: 0x00000000 response 0xffffffc2
[   61.855812] amdgpu: [powerplay] failed send message: GetMaxDpmFreq (31)  param: 0x00000000 response 0xffffffc2
[   64.330326] amdgpu: [powerplay] failed send message: GetMaxDpmFreq (31)  param: 0x00020000 response 0xffffffc2
[   66.805799] amdgpu: [powerplay] failed send message: GetMaxDpmFreq (31)  param: 0x00020000 response 0xffffffc2
[   69.018673] amdgpu: [powerplay] failed send message: NumOfDisplays (64)  param: 0x00000002 response 0xffffffc2
[   71.231672] amdgpu: [powerplay] failed send message: NumOfDisplays (64)  param: 0x00000002 response 0xffffffc2
[   73.710662] amdgpu: [powerplay] failed send message: GetMaxDpmFreq (31)  param: 0x00000000 response 0xffffffc2
[   76.190010] amdgpu: [powerplay] failed send message: GetMaxDpmFreq (31)  param: 0x00000000 response 0xffffffc2
[   78.657504] amdgpu: [powerplay] failed send message: GetMaxDpmFreq (31)  param: 0x00020000 response 0xffffffc2
[   81.138088] amdgpu: [powerplay] failed send message: GetMaxDpmFreq (31)  param: 0x00020000 response 0xffffffc2
[   83.617616] amdgpu: [powerplay] failed send message: GetMaxDpmFreq (31)  param: 0x00000000 response 0xffffffc2
[   86.096937] amdgpu: [powerplay] failed send message: GetMaxDpmFreq (31)  param: 0x00000000 response 0xffffffc2
[   88.574473] amdgpu: [powerplay] failed send message: GetMaxDpmFreq (31)  param: 0x00020000 response 0xffffffc2
[   91.051993] amdgpu: [powerplay] failed send message: GetMaxDpmFreq (31)  param: 0x00020000 response 0xffffffc2

The remaining 30 to 40% of the boots run smooth and I end up on the desktop after a few seconds. The above boot also ends up on the desktop eventually (after around 45 seconds), but for example launching steam gives another slew of powerplay fails with the delays that go along with that.

BUT: Even when things go smoothly, Rise of the Tomb Raider gives me 30fps in the benchmark (already had that with Ubuntu 19.10), with no powerplay fails. On the contrary an 18.04 Ubuntu with Mesa 19.3 from the Kisak PPA and the 5.3 distro kernel gives me 140+ fps - about 5 times more.

Bottom line: If things change not dramatically in the next few weeks an RX 5700 will be unusable in a fresh Ubuntu 20.04. By that time the GPU will have been on the market for nearly a year. AMD - why can't you get your shit together?

Don't use 5.3 or 5.4, seriously. If you want improved experience, use 5.5 or 5.6. Powerplay bugs are fixed in only in 5.6 currently.

Last edited by Shmerl on 16 February 2020 at 9:57 am UTC

Or 5.3...

I briefly had the 5.5 kernel running the 20.04 - also gave me a powerplay failure. Anyway, the bottom line remains: If someone installs a brand new Ubuntu 20.04 in April (or an equally popular spin-off like Mint) his RX 5700 experience will be most likely dreadful. And the 5.5+ won't land in Ubuntu proper before Ubuntu 21.04.
Anyway, I'll give 5.6 a try once it draws closer to stable.

Last edited by Tuxee on 16 February 2020 at 10:03 am UTC

Those who use AMD know to use latest kernels, so don't worry about Ubuntu. It's never good for that, Navi or not.

Last edited by Shmerl on 16 February 2020 at 10:06 am UTC

That's why I recommend NVidia to each and everyone. Because with AMD you will inevitably end up as the Windows user's cliché of the "always tinkering Linux nerd without ever getting anywhere". Correct?
Point is: Someone wants to switch to Linux has a decent desktop computer set up and will head for the most popular distros Ubuntu or Mint. And to be fair, there is a reason why they are popular. They will get the recommendation to stick to the nice LTS versions and now someone comes along and tells them to become "nerds" and do all the crap they heard about Linux in their Windows era: Distro hopping, reading logs, system tinkering, perhaps some kernel compiling. And all because AMD can't get its shit together.