Nouveau Linux 4.8 + Mesa 12.1-dev vs. NVIDIA OpenGL Performance

[imirkin]

It's a little unfortunate that these results are carefully (carelessly?) done so as present nouveau in the worst light possible.

The real issue is that on all but the GTX 680 board, memory reclocking completely fails. So you have a high core speed and the original lowest memory clock. This yields totally crap performance (as can be seen).

One would likely get better results by first switching to a middle perf level (usually called "0a" on such GPUs), which should allow the memory clock change without the voltage change, and then try to flip to 0f. The memory reclock will fail again, but at least you'll have a middling memory clock instead of the absolute lowest.

But realistically, any user seriously considering nouveau for gaming would never use upstream. They'd use a tree with Karol's reclocking patches, where every Kepler GPU should behave roughly as the GTX 680 one does - about 60-80% of blob performance, depending on the application.

[molecule-eye]

And this is why I refuse to buy nvidia hardware even if it performs better. For my purposes, I can get buy with fluid performance using open source drivers and moderate AMD hardware (an RX 460), so there's absolutely no point in giving my money to nvidia. And if you like testing/using/ Wayland, there's another reason to stay RED.

[dungeon]

But realistically, any user seriously considering nouveau for gaming would never use upstream.

He, he, but really realistically articles speak to most people so to Average Joe, who typically uses upstream kernels which are not even up do date always

[cj.wijtmans]

This is why you should believe nvidia when they come up with something else than GBM.

[karolherbst]

The real issue is that on all but the GTX 680 board, memory reclocking completely fails. So you have a high core speed and the original lowest memory clock. This yields totally crap performance (as can be seen).

actually engine reclocking fails due to voltage mismatches

[smitty3268]

This is why you should believe nvidia when they come up with something else than GBM.

Because they don't publish the specs to change power states on their hardware?

[AdamOne]

I've been running Linux exclusively for almost 15 years. And for all that time NVidia has provided fast and working drivers. I have a hard time calling that "heinous".

I wrote "Nvidias perspective on OS", you ought to know what thats short for, after 15 years

[trek]

I don't understand why Nvidia doesn't offer (aside closed) open power management to the Linux kernel.

I've got the idea that Nvidia has superior software (but not hardware) and they try defend their advantage improving the closed source driver and not helping open source knowledge sharing

[pal666]

Is there a proper explanation for that?

i don't know, maybe because 20 developers is more than 2? (numbers are arbitrary)

[pal666]

I've been running Linux exclusively for almost 15 years. And for all that time NVidia has provided fast and working drivers. I have a hard time calling that "heinous".

you have been running nvidia-blessed kind of linux in nvidia-blessed way with some strange definition of "working". you could just as well run windows. https://www.youtube.com/watch?v=iYWzMvlj2RQ