Windows 10 Radeon Software vs. AMDGPU On Ubuntu Linux

[theriddick]

I tried the PRO driver out before but didn't work on my 16.04 install, guessing you need to regress the kernel back to 4.2 for it to work.

[bug77]

In my opinion, if an application fits (when running) into a 32-bit address space, and does not benefit from 64-bit integer arithmetic, there is nothing wrong in compiling it for a 32-bit target. For example, small "streaming" Linux programs such as 'cat', 'tr', 'tail', or small programs such as 'cron' and 'sleep', fall into this category.

Well,. it does force you to install a bunch of 32bit libraries which may come with vulnerabilities on their own. There are worse things in life, true, but this still qualifies as an unnecessary annoyance.

[Taupi]

In real life, we play (games that can require power from CPU/GPU stuffs and can be evaluated on windows and linux -and even mac-):
- on steam: Civ V/beyond Earth, Borderland 2, Medieval: Chivalry, CS:GO, Ark:Survival, Portal2, L4D2, XCOM2, Bioshock:infinite, Middle EArth:Shadow of Mordor, Mediaval: Total War 2 ...
- with GOG games: Witcher2, Pillar of eternity, Divinity:OS, Wasteland 2, Mount&Blade:warband, SubLevelZero, SuperHot, FireWatch
- with Unreal Tournament 4: https://www.epicgames.com/unrealtour...l=1#post191678

I hope that Mickael will test that kind of games one day ^^

[Mike Frett]

duby229: phoronix tests have always been like this. There is obviously not much real-world value in these benchmarks, especially when it comes to Linux vs Windows gaming. But it is still enough to indicate trends or reveal bigger problems.
But this is still the #1 place to go for linux benchmark results and therefore influential. When people consider Linux and read about drawbacks like poor gaming capabilities, they search the web for confirmation. They lose their interest and developers/publishers don't get interested at all. So imho, good results are important to see here

It's unfortunate that Benchmarks are more important to people than Privacy and Security.

[juno]

I hope that Mickael will test that kind of games one day ^^

He does only test what can be tested automated from the command line.
Also, it's not only the selection of games that makes the tests less meaningful. FPS numbers don't say anything at all about smoothness and feeling. A better metric are frametimes and even better is a deeper analysis, but that's nothing I am remotely asking for here.

It's unfortunate that Benchmarks are more important to people than Privacy and Security.

You don't have to be that bitter when the world is not that black/white. You can still have privacy and security, use Linux in your everyday life and play games on Windows.

[bug77]

It's unfortunate that Benchmarks are more important to people than Privacy and Security.

Why? Has your video card been spying on you? Emptied your bank account?

[Passso]

So majority of people must get broken steam after they install it because it FORCES itself to run these bad OLD libraries. Seems rather piss poor solution if you ask me!

That is the case for 75% of installations yes... and 64 bits OS makes things even more worse...

[Passso]

Most people, the vast majority, just use the igp.

Please read the Steam statistics about hardware first. Then write.

[theriddick]

Talos has deeper framerate analysis data. It would be nice if a curve chart was possible showing how consistent framerate was. I know that some games are shockers yet will try and tell you they get high fps but the stutter is terrible. (Warthunder on the AMD drivers for example, XCOM2 is another).

[rabcor]

I am impressed... Maybe I'll go AMD next... (Although the scores are looking less impressive when you compare them to Nvidia's performance; although there is a certain lack of on-par samples, e.g. the Fury generally exhibits performance similar to 980, Fury X to 980-Ti, and I imagine the 285 would be somewhere around the range of 960, although it seemed to be performing noticably worse than even a 950)

But at least this tells me that AMD are indeed doing their best here, and that is enough for me. Only way to go now is up.

Please read the Steam statistics about hardware first. Then write.

The vast majority of people do use igp. It's the vast majority of PC gamers that don't.

In my opinion, if an application fits (when running) into a 32-bit address space, and does not benefit from 64-bit integer arithmetic, there is nothing wrong in compiling it for a 32-bit target. For example, small "streaming" Linux programs such as 'cat', 'tr', 'tail', or small programs such as 'cron' and 'sleep', fall into this category.

There isn't, but steam is a slow as fuck performing application, and it's easy to point the fingers at shitty 32-bit code rather than simply just the fact that the code must obviously be shitty to begin with. Volvo have been lazy about this for years and years on end.