Fedora Project Leader Calls Out NVIDIA Over Their Proprietary Linux Drivers

[stargeizer]

Dunno about openGL, but if you ask any developer that aren't bound by an NDA (i do, even tho i'm not working anymore with computers, retirement is such a glorious thing, and this is not 4chan, so no more comments from me) maybe they will tell some similar stories about Nvidia engineers. Great people BTW!.

And, don't know about the post you mention, google gives me nothing...

EDIT: OK! is related to this??

https://www.geeks3d.com/forums/index.php?topic=3588.0

Well, this kinda confirm my point i guess... But i'm not going to lie to you: I would have loved to work for Valve. Born in the wrong continent i guess

And birdie, no coder that have seen confidential code and signed NDA's can give you proof. That's the whole point of such annoyances and that's why anyone will only find anecdotes. But feel free to discard what i've written in these forums. After all, this is Phoronix forums. Be happy at least this is not 4chan, but sometimes i wonder if i should do better things with my remaining time

[birdie]

I have to say some of the stuff you wrote here sounds really familiar. Is it from that 2014 blog post of a Valve employee that discussed the OpenGL driver quality of different vendors?

Some of this is true but the rest of the post looks like crazy BS.

Many years ago, I briefly worked at NVIDIA on the DirectX driver team (internship). This is Vista era, when a lot of people were busy with the DX10 transition, the hardware transition, and the OS/driver model transition. My job was to get games that were broken on Vista, dismantle them from the driver level, and figure out why they were broken. While I am not at all an expert on driver matters (and actually sucked at my job, to be honest), I did learn a lot about what games look like from the perspective of a driver and kernel.
The first lesson is: Nearly every game ships broken. We're talking major AAA titles from vendors who are everyday names in the industry. In some cases, we're talking about blatant violations of API rules - one D3D9 game never even called BeginFrame/EndFrame. Some are mistakes or oversights - one shipped bad shaders that heavily impacted performance on NV drivers. These things were day to day occurrences that went into a bug tracker. Then somebody would go in, find out what the game screwed up, and patch the driver to deal with it. There are lots of optional patches already in the driver that are simply toggled on or off as per-game settings, and then hacks that are more specific to games - up to and including total replacement of the shipping shaders with custom versions by the driver team. Ever wondered why nearly every major game release is accompanied by a matching driver release from AMD and/or NVIDIA? There you go.

The second lesson: The driver is gigantic. Think 1-2 million lines of code dealing with the hardware abstraction layers, plus another million per API supported. The backing function for Clear in D3D 9 was close to a thousand lines of just logic dealing with how exactly to respond to the command. It'd then call out to the correct function to actually modify the buffer in question. The level of complexity internally is enormous and winding, and even inside the driver code it can be tricky to work out how exactly you get to the fast-path behaviors. Additionally the APIs don't do a great job of matching the hardware, which means that even in the best cases the driver is covering up for a LOT of things you don't know about. There are many, many shadow operations and shadow copies of things down there.

The third lesson: It's unthreadable. The IHVs sat down starting from maybe circa 2005, and built tons of multithreading into the driver internally. They had some of the best kernel/driver engineers in the world to do it, and literally thousands of full blown real world test cases. They squeezed that system dry, and within the existing drivers and APIs it is impossible to get more than trivial gains out of any application side multithreading. If Futuremark can only get 5% in a trivial test case, the rest of us have no chance.

[Linuxxx]

Nearly every game ships broken. We're talking major AAA titles from vendors who are everyday names in the industry.

Thankfully, DXVK is able to fix alot of Direct3D API misusage mistakes over on the Vulkan side!

Recent example: The Evil Within

https://github.com/doitsujin/dxvk/pull/2522

[Nozo]

Welp.

You're a user of the OS who most people know nothing about which has stagnated on the desktop for the past 30 years and never broke out of ~2% desktop market share.

Have fucking modesty. NVIDIA, nor any other proprietary company owes nothing to you and your particular OS. Show me a single reason why Linux is better than Haiku OS, FreeBSD, NetBSD, Open BSD, Solaris, QNX and a ton of other OSes. Why the fuck NVIDIA should support Linux more than they already do?

Anyone ever say that?
Also those operating systems are not the best examples in terms of popularity.

Nothing will change in terms of Linux popularity and acceptable between both users and ISVs before this gets fixed/implemented/whatever once and for all.

Says the one who mentions operating systems that all together are not capable of exceeding 1% market share, even though some of them have stable APIs/ABIs, simply because nobody cares about Haiku, nobody cares about BSD, nobody cares about Solaris/Illumos. and they are operating systems that have been around for decades as well. No one knows them, and the companies that do parasitize them at best. Instead of blaming ghosts and repeating what others repeat on the internet, the real problems and obstacles that prevent all these operating systems, not just Linux, from gaining mass adoption should be located.

[clockwork]

And you continue with this BS which I've already addressed three times and others as well.

You didn't address shit lol. I explained to you how Nvidia uses game profiles on the Windows drivers and you didn't respond. Your comment demonstrated your lack of understanding of the actual drivers. And the "other" user basically stated the same thing I did, and pointed to the Nvidia Profile inspector tool. If there are no profiles WHAT DOES THAT TOOL DO?! lmao. You argue like a politician. Poorly. Moving on from you lol.

[birdie]

You didn't address shit lol. I explained to you how Nvidia uses game profiles on the Windows drivers and you didn't respond. Your comment demonstrated your lack of understanding of the actual drivers. And the "other" user basically stated the same thing I did, and pointed to the Nvidia Profile inspector tool. If there are no profiles WHAT DOES THAT TOOL DO?! lmao. You argue like a politician. Poorly. Moving on from you lol.

So do AMD and Intel, so what? Per game optimizations have existed for as long as GPUs existed. The rest of your post was complete and utter bullocks and I asked for proofs and here you are with the 4th post on the matter with absolutely nothing for it. I won't reply to any of your further posts from now on. Tired of dealing with windbags, conspirologists and buffoons.

[jacob]

For folks starting out new, using Intel or AMD might make more sense. However Nvidia has considerable marketshare and some users cannot easily move away (CUDA for example) and it would be good to have all the major players working together instead of one vendor standing out and stalling progress on several common projects that adversely affects even users who use Intel or AMD.

It would be good, however when one of them refuses to be compatible and interoperabile, it should be made their own loss. You are right about CUDA, but if I relied on CUDA-based software, the easiest by far would be to simply run it on Windows.

[RahulSundaram]

You are right about CUDA, but if I relied on CUDA-based software, the easiest by far would be to simply run it on Windows.

CUDA is used in a lot of cases where Linux servers actually do make better sense for overall performance however. So it is not an easy tradeoff.

[StillStuckOnSI]

but if I relied on CUDA-based software, the easiest by far would be to simply run it on Windows.

A large percentage of software using CUDA only runs on Linux.

[drakonas777]

Oh, I see our usual NVIDIA fanboys are defending beloved company yet again. Leave them alone I think, guys probably hadn't a sleep in two days fighting criticism of NVIDIA LOL

BTW I've changes GTX 1060 to RX 6600 in my main Linux desktop and all problems gone. No more flickering during remote desktop sessions etc. Best decision ever.