FreeBSD: A Faster Platform For Linux Gaming Than Linux?

[kraftman]

It was the same hardware between Linux and FreeBSD...

Great, but didn't you know you benchmarked compiz against kwin, so this makes entire comparison useless? What a crap.

[kraftman]

I understand these are out of the box experiences from each OS. In that regards the benchmarks are relevant in my opinion.

They're not relevant, because of the obvious reasons:

- Ubuntu is a Linux distribution not Linux. It was showed Ubuntu has the slowest 3D performance with nvidia blob compared to other Linux distributions

- he wanted to check this: "When a system call is called by the Linux binary, the trap code dereferences the system call function pointer off the proc structure, and gets the Linux, not the FreeBSD, system call entry points...", but he just measured kwin vs compiz

If the tests were about Ubuntu and some BSD distro 3D performance then everything will be fine, but they're not about this. He could at least chose Kubuntu to make the tests more relevant...

[DMJC]

10% is a massive difference

Wine needs to get benchmarked on this. 10% is a massive performance difference, as we see with some of the early benchmarks 16fps vs 26fps those 10 fps are enough to make the game go from unplayable to smooth/playable. It would be interesting to see if wine is more competitive on BSD than on Linux when it comes to getting near the performance of windows.

[yossarianuk]

Well Ubuntu benchmarks slower at all games than Arch for me.

I have stopped using Ubuntu as it was slower that pretty much every distro I have tried (on my machine) at all games, flash, video playback ,etc.

Was composting enabled in both systems?

[Shining Arcanine]

I would like to see FreeBSD against Gentoo Linux, which I assume that should be one of the best performance Linux distribution.

How about FreeBSD versus PC-BSD versus Gentoo/FreeBSD versus Gentoo Linux benchmarks?

I imagine that much of the performance advantage Gentoo would have is from having a newer toolchain with more specific optimizations, so adding Gentoo/FreeBSD into the mix would nullify that advantage.

FreeBSD vs Linux or Unity vs KDE?

Here's a comparison with some previous phoronix benchmarks of Desktop Environments.





Notice the performance difference between KDE and Unity when using the NVidia driver and the similarities between the FreeBSD vs Linux results. I know enough. You just benchmarked Unity vs KDE again. And you prove again that the Nvidia driver performs best with KDE.

I wish Michael had done this comparison before he published this. He just got slashdotted and I am sure few people are reading the forums to find out the truth. It is good advertising for PC-BSD though. I just wish it was clear that PC-BSD by default is faster than Ubuntu by default.

http://games.slashdot.org/story/11/09/08/0437248/Linux-3D-Games-Run-Faster-On-PC-BSD

I find I get smoother performance on Linux using the low-latency version of the kernel.
(

Your regular kernel likely does not have preemption support. Using the "low-latency" versus implies using a realtime kernel, that adds overhead, but no real benefit unless you want to control an industrial laser. Having a preemption based kernel is the best of both worlds. You can check what your stock kernel is by doing uname -v. Here is it on my system:

$ uname -v
#2 SMP PREEMPT Wed Aug 31 08:51:10 EDT 2011

[Gunss]

But you think that a preemptive kernel should help in game performance?


At least in multimedia it helps a lot!

[russofris]

I wish Michael had done this comparison before he published this. He just got slashdotted and I am sure few people are reading the forums to find out the truth. It is good advertising for PC-BSD though. I just wish it was clear that PC-BSD by default is faster than Ubuntu by default.

It's pretty clear that this caparison was not made specifically because "getting slashdotted" was the intended and desired outcome. Who needs good content when you can throw up a controversial headline, fill an article with fluff, and spam aggregators with links to drive up ad revenue?

Quite sad.

[Nahemah]

Fanboys revolt!

Quick! Revolt! It MUST be propaganda!! Linux fails at being Linux? IMPOSSIBLE!

Yeah, BSD kicks your ubuntu ass. Deal with it.

Oh don't worry Linux users, you still beat Win98.

[tuke81]

Please read my post again. png. Yes I have read that thread.

Ah. My bad, Yes you are right with png that sure look like you told.

[Desti]

cpu: SAME
ram: 1GB less then Linux on i386, so Linux should be faster here
motherboard/chipset: SAME (MS-7642)
display driver: SAME VERSION
graphics card: SAME
compiler: NEWER on Linux, so Linux should be faster here
xorg sever: NEWER on Linux, so Linux should be faster here
fs: EXT4 is nowhere near ZFS features, and is probably more 'heavy' then EXT4, so Linux should be faster here

KDE vs. UNITY: both should not be used during that test, it should be plain twm or openbox with all needles services disabled on both systems.



If you would want to setup that from FreeBSD, it would be quite more work then with Ubuntu, but PC-BSD has graphical installer as Ubuntu and the amount of work needed to install con configure Ubuntu is the same as the amount of work needed to setup PC-BSD.

The size of the memory doesn't matter for any of this benchmarks, none of them needs more than 2 GiB. But the speed of the memory does matter, but that's nowhere mentioned, so I guess he even don't know it.