AMD Radeon RX 5500 XT Linux Performance

[coder]

I do acknowledge that you do and this is a subjective matter,

That's a point of disagreement, then. Because I think it looks objectively better. There will always be people with eclectic tastes, but I'm talking about the average person.

That said, there certainly could be issues with how your device implemented the process. Ideally, we'd be able to watch HFR source material and just skip any form of temporal interpolation, entirely.

edit: I have yet to see the Gemini Man, however since it's in 3D I can see that 120Hz helps there since 3D needs higher framerates due to how the 3D technology works.

The weird thing is that they filmed it in 3D @ 120 Hz. However, technical limitations in current theaters meant that it could only be shown in either 4k @ 60 Hz or 2k @ 120 Hz. Only a handlful of theaters in the US showed the 120 Hz version.

I see however that The Times did not enjoy the HFR:

Wow. What a loser. I'll bet he watched the movie just expecting to hate it. I would not trust other reviews buy this guy, since he clearly has trouble with objectivity.

It was a solid action film, IMO. Nothing too special, but I found the reverse-aging to be good enough that I wouldn't have paid it any mind, if I hadn't been aware of what they actually did. There were a couple points when it looked a touch weird, like maybe an unnatural micro-expression, during complex emotional sequences. Ironically, had it not been HFR, such glitches would've been much less noticeable.

[Linuxxx]

coder
Keep enjoying your artifacts, then!

Also, to just assume that the interpolation I suggested would result in a "blurry mess" without having at least tried it out beforehand - well, let me just quote you:
>> WOW! What a loser... <<

[bitnick]

Because I think it looks objectively better.

The I think part is really the definition of something subjective. I.e. it's an opinion (it's something subjective) and not a statement of fact (something objective) that it looks better. Just saying... Myself, I haven't really formed an opinion about HFR yet. Not enough experience.

[F.Ultra]

That's a point of disagreement, then. Because I think it looks objectively better. There will always be people with eclectic tastes, but I'm talking about the average person.

That said, there certainly could be issues with how your device implemented the process. Ideally, we'd be able to watch HFR source material and just skip any form of temporal interpolation, entirely.

No, I have one of the absolute top of the line UHD TVs (which have a 24fps native mode) at home (been a home cinema nerd for many many years)so the problem is not the implementation. The problem (that is equally problematic for every single algorithm) is that #1 new frames have to be created by the existing frames since there does not exist any additional temporal information in the existing frames and #2 that there just does not exist a smooth way to play 24fps at 60fps since there are no common factors. Don't know where you are from but a major problem with 24fps on TVs have been the NTSC system since it uses 60Hz and thus 24fps had to be converted to 60Hz by 3:2 pull down which makes camera movement jerky which is something that was never the case here in Europe since we simply just played it 4% faster at 25fps to match our 50Hz PAL.

[aht0]

IMO, it's good that they managed a 128-bit card that can hang with their previous-generation 256-bit cards, but the value is not there. RX 570 performs nearly as well, and if you look at the current sale prices, in particular, its performance per $ is way better. Plus, Polaris has no caveats around support for GPU compute.

Over time, RX 5500 cards should come down in price, hopefully to the RX 560's bracket. However, at least until RX 570 stocks run dry, the 5500's will be challenging to recommend.

Hmm, yeah, RX580 and especially RX590 seem both like a better deal for now. More memory bandwith and more use on higher resolutions.. Price? At low hundreds bit more or bit less ain't that important.

[arQon]

Zero RPM does not work in Linux.

Just FYI, that's not an "in Linux" problem. Reducing fan speed is also broken in Windows: sometimes it might work if you leave it for 5-10 mins, other times you're looking at a reboot.
Like almost everything else to do with AMD's drivers, it's broken, it's been broken for literally years, and it will doubtless STAY broken for years longer.

AMD's drivers are absolute garbage. (Sorry bridgman). Every year they put a new coat of paint on them and add a stupid new gimmick to them (this year's was a web browser - because what you really need on a Windows machine is a process running with SYSTEM rights and an attack surface the size of Texas. What could possibly go wrong with THAT brilliant idea?), but the core functionality hasn't worked properly in longer than anyone can remember.

The drivers are so bad that they're enough of a reason to buy an nvidia card instead, despite the nontrivially-worse perf per dollar and even with full knowledge of nvidia's terrible past behavior. It breaks my heart.

Oh - and let's not forget the ads built into the drivers, which are more obnoxious than a Facebook page these days.
The driver package is *over a GIG now* because of all this crap - but basic functionality still just doesn't work, and at this point it's obvious that RTG just doesn't even care any more. It's insane.

[arQon]

A web browser is not in Mesa, is not in radeon/amdgpu kernel modules and is not in the amdgpu-pro-opencl package.

The Linux drivers are much better, sure. I was originally just pointing out that you're mistaken thinking that your fan control problems were caused by the Linux drivers: they're not, it's a firmware / other shared piece.

re the browser etc, good to know - except user-reachable data is more important than the OS anyway, and it's still an utterly retarded thing to try to hack into a *driver*, ffs.

[aht0]

Just FYI, that's not an "in Linux" problem. Reducing fan speed is also broken in Windows: sometimes it might work if you leave it for 5-10 mins, other times you're looking at a reboot.
Like almost everything else to do with AMD's drivers, it's broken, it's been broken for literally years, and it will doubtless STAY broken for years longer.

Have you tested on multiple different cards or you are having an issue just with your own? Following what's written in reddit, cards seem to differ in quality from vendor to vendor (Sapphire's being the best).

AMD's drivers are absolute garbage. (Sorry bridgman). Every year they put a new coat of paint on them and add a stupid new gimmick to them (this year's was a web browser - because what you really need on a Windows machine is a process running with SYSTEM rights and an attack surface the size of Texas. What could possibly go wrong with THAT brilliant idea?), but the core functionality hasn't worked properly in longer than anyone can remember.

Ain't that much different from systemd then (network connected, PID1, nice wide attack surface), tho I don't think that particular process is running under SYSTEM privileges on Windows.

The drivers are so bad that they're enough of a reason to buy an nvidia card instead, despite the nontrivially-worse perf per dollar and even with full knowledge of nvidia's terrible past behavior. It breaks my heart.

Nvidia's performance sure is better, I had recently chance to use Nvidia and Radeon simultaneously and I think latter offered better picture quality tho. Comments?
After recent Radeon-loaner, I am considering buying new RX590 or 580 for myself (new because I don't want cards whored by cryptominers). Can you be more specific about the issues you've had?

Oh - and let's not forget the ads built into the drivers, which are more obnoxious than a Facebook page these days.
The driver package is *over a GIG now* because of all this crap - but basic functionality still just doesn't work, and at this point it's obvious that RTG just doesn't even care any more. It's insane.

Nvidia drivers are also telemetry-enabled (sending routinely data out).

[arQon]

The bigger problem is simply this: AMD's drivers have been unequivocably broken for literally years now. Things don't get into that sort of state as a result of one-off crunches etc, but because of systemic problems like a corporate culture of "bugs aren't important" and / or a technology team that is woefully mismanaged or under the thumb of a marketing department that thinks a sufficiently red and black color scheme is the only thing a piece of software actually needs, regardless of how much of its core functionality doesn't even work.

AMD used to also release "sane" versions of the drivers, without all this "Gaming Evolved v2" crap in them. Those versions may still exist, but if so they've been hidden even further out of sight lately. So now, if you want your hardware to be usable at all (on Windows, at least) the only option you have is this "gamer-iffic" gimmick-laden monstrosity.

[aht0]

AMD used to also release "sane" versions of the drivers, without all this "Gaming Evolved v2" crap in them. Those versions may still exist, but if so they've been hidden even further out of sight lately. So now, if you want your hardware to be usable at all (on Windows, at least) the only option you have is this "gamer-iffic" gimmick-laden monstrosity.

Have you tried to extract "driver pieces" from installer and install just those without the extraneous bloat? I've done so with Nvidia's - unpacked whole executable with 7z, erased folders useless to me (Shield drivers, Nvidia HDMI Sound drivers, Geforce Experience, Telemetry etc) and allowed Windows itself to "search for drivers" from that root folder of the extracted driver. After Windows has found driver for primary GPU, rest that unpacked folder I would simply delete.

I kind of remember that I also used to unpack Catalyst drivers (chipset/cpu/usb3) and incorporate these into Windows iso. So in theory same approach should work for Radeon drivers.