Ubuntu 13.04 Linux Can Outperform Apple OS X 10.8.3

[enfocomp]

Thanks for the benchmarks. Good to see some 3D games benchmarked & pushing slightly ahead of Mac OS X. Ubuntu 13.04 is performing better then 12.04 LTS in the vast majority of benchmarks and it's progressing a lot quicker then Mac OS X.

Keep up the good work Micheal!

[smitty3268]

Phoronix used to have ACTUALLY relevant tests like h264 encoding

WHAT HAPPENED?

Michael started focusing more on the test framework than the actual tests.

[dee.]

Fuck results.
Apple fanboyism >> technical and perf superiority of Linux.

And Apple is soooo much better. It made me give up my typewriter.

Miguel, is that you?

[Pallidus]

Phoronix used to have ACTUALLY relevant tests like h264 encoding

WHAT HAPPENED?

I would like to see more useful tests:
- compile times (same compiler)
- SunSpider & other JavaScript benchmarks (of the same codebase of-course)
- Unigine & other modern 3D engines running at high resolutions.

And why?
- Add Ubuntu's tests compiled under LLVM

[chrisb]

What an absolute useless pile of dung comparison set. No one at Apple gives a rat's ass about Phoronix's Test Suite, never mind comparing start up times of HFS+ and Ext4.

Seriously, this is actually embarrassing for Linux.

There was no comparison of start up times, the comparison was of archive unpack times. I doubt there is much embarrassment at performing significantly better.

Having said that, you are probably right that people at Apple aren't so concerned about absolute performance figures, since very few Mac buyers will make a purchase decision on that basis.

[snadrus]

Better tests

I would like to see more useful tests:
- compile times (same compiler)
- SunSpider & other JavaScript benchmarks (of the same codebase of-course)
- Unigine & other modern 3D engines running at high resolutions.

And why?
- Add Ubuntu's tests compiled under LLVM

[Marc Driftmeyer]

What an absolute useless pile of dung comparison set. No one at Apple gives a rat's ass about Phoronix's Test Suite, never mind comparing start up times of HFS+ and Ext4.

Seriously, this is actually embarrassing for Linux.

[frign]

Some maths

Which is not hard, because OS X is even more a bloatware than Ubuntu is.

And because bloat(OS X) > bloat(Ubuntu) and we have the relation speed(OS) = 1 / bloat(OS), this test yields to the expected results.

Now seriously: Apple has really let go with its Desktop Operating System. It is a joke and starting with Snow Leopard was one of the main reasons I switched to Gentoo.

[ltcommander.data]

I'm not sure OS X GPU benchmarks of Ironlake are meaningful. Ironlake was slower than the previous MacBook Pro IGP, the 9400M, so Apple never implemented it as a standalone IGP for the 13" MacBook Pro, Mac Mini or iMac unlike the 9400M or the later HD3000/HD4000. It was only ever implemented in a switchable configuration for power savings where the OS automatically switches to the discrete GPU under any reasonable GPU load so the Ironlake GPU driver isn't likely performance optimized. Apple doesn't even acknowledge the existence of Ironlake as a targetable GPU to developers unlike the Intel GMA 950, X3000, HD3000, HD400 and nVidia 9400M and 320M which were available in standalone configurations so are going to need driver development effort.

https://developer.apple.com/graphicsimaging/opengl/capabilities/GLInfo_1075.html

https://developer.apple.com/graphicsimaging/opengl/capabilities/index.html

[blackout23]

What about power consumption?

What about level of gayness? OSX outperforms anything in this benchmark.