Windows 10 vs. Linux - Intel UHD Graphics OpenGL Performance

[phoronix]

Phoronix: Windows 10 vs. Linux - Intel UHD Graphics OpenGL Performance

Over the past week I have carried out some Radeon and NVIDIA Windows 10 vs. Ubuntu Linux benchmarks. While not quite as interesting as those discrete GPU comparisons, while having the Windows 10 Pro x64 Fall Creator's Update around and testing from the Core i7 8700K, I also ran some Windows vs. Linux tests for the integrated UHD Graphics.

http://www.phoronix.com/vr.php?view=25602

[microcode]

I wonder what makes the difference for Clear Linux. I think it would make sense on other distros to compile the Intel DRI in a fashion which is optimized for CPUs which actually include a compatible Intel accelerator, even on general-purpose distros.

[treba]

I guess some of the performance difference still lies in the gnome-shell wayland implementation. As of now, it doesn't use hardware planes for all fullscreen rendering, causing some overhead. I experience this myself in games where my (low end) hardware is at its limits. Switching to xorg makes quite a difference there.

I'm not sure but maybe the reason why we don't see that in xonotic is that it's rather cpu bottlenecked while the first benchmarks maybe have a bandwidth problem, which gets worse when there're extra copies going on
I hope Daniel van Vugt or Jadahl will have time to look into it.

[LinuxID10T]

Actually, everyone else is wrong. Ubuntu is slower because the performance governor is set to powersave. In Clear Linux it is set to performance.

[lucasbekker]

Could the modesetting driver also be to blame for the lower performance compaired to windows?

[LinuxID10T]

Could the modesetting driver also be to blame for the lower performance compaired to windows?

Modesetting is just the 2D driver. Shouldn't affect 3D performance.

[Guest]

Modesetting is just the 2D driver. Shouldn't affect 3D performance.

Doesn't modesetting force 2D rendering to be ran through 3D?

[ext73]

Michael do You know that kernels 4.14 and 4.15 have a mega bug [regression] in Intel i915 drivers ?!?! Somehow, for now, the Intel developers 'are not eager' to fix it
- for example:

https://github.com/NixOS/nixpkgs/issues/31999

https://forum.manjaro.org/t/poor-opengl-performance-on-linux-4-14/35453

https://bbs.archlinux.org/viewtopic.php?id=232354

Regards

[Vistaus]

Michael do You know that kernels 4.14 and 4.15 have a mega bug [regression] in Intel i915 drivers ?!?! Somehow, for now, the Intel developers 'are not eager' to fix it
- for example:

https://github.com/NixOS/nixpkgs/issues/31999

https://forum.manjaro.org/t/poor-opengl-performance-on-linux-4-14/35453

https://bbs.archlinux.org/viewtopic.php?id=232354

Regards

Is that true? 'Cause when I mentioned 4.14 on OMG! and said there were serious Intel regressions, I only got a couple of replies stating that nothing was wrong at most people's machines and that my hardware was to blame, etc.

[treba]

Actually, everyone else is wrong. Ubuntu is slower because the performance governor is set to powersave. In Clear Linux it is set to performance.

I doubt that. It's not the classical version of powersafe, where the cpu was just throttled to minimum speed. It's the intel_pstate governor. Sure there's some overhead of the powersaving mode, but it shouldn't make a huge difference, as we can see from the xonotic benchmarks which probably are more cpu-limited as the unigine benchmarks.

BTW. funny use of language "Actually, everyone else is wrong."