Apple Mac OS X 10.7 DP2 Battles Ubuntu 10.10

[kraftman]

I thought GCC 4.2.1 is deprecated in Lion and LLVM is now the default for Lion. So perhaps it'd makes sense to do the benchmarks using LLVM across Snow Leopard, Lion, and Ubuntu.

Ubuntu doesn't use llvm for default. Llvm can't even compile Linux. It's logical GCC which is shipped with Ubuntu should be used.

[Qaridarium]

Funny, as when I publish OS benchmarks that use the upstream packaged compiler for each OS, people complain loudly as well that it's then an inaccurate comparison.

yes thats funny.. but i never make a point out of this,.

means stock stuff for every os is valid and newest shit vor every os is also valid..

but why damange a system by hand? only to get a result for apple fanboys?

[Qaridarium]

Well, you basically had two choices here, either choose the same compiler version or use default stock compiler. With the first choice people complain as it in fact is unrealistic scenario and with the second choice people would complain that it is unfair comparing different versions etc. I'd say stick to your "default" motto and show a finger to people telling you otherwise...

default stock is the only valid way! no one hurt his system only because apple do not wana use GPLv3 stuff.

but really no one get payed from apple to use old software in linux.

[ltcommander.data]

default stock is the only valid way! no one hurt his system only because apple do not wana use GPLv3 stuff.

but really no one get payed from apple to use old software in linux.

It really depends on what the point of the comparison is. If its to compare general end user experience, then stick to default options for each OS. If its to isolate the OS contribution to software performance, then the same software needs to be used on each OS. If its to look at the max potential of each OS, then the latest released or even unreleased versions of each software available for each OS should be used even if they are not the same between OS. All provide useful but different information.

[Qaridarium]

It really depends on what the point of the comparison is. If its to compare general end user experience, then stick to default options for each OS. If its to isolate the OS contribution to software performance, then the same software needs to be used on each OS. If its to look at the max potential of each OS, then the latest released or even unreleased versions of each software available for each OS should be used even if they are not the same between OS. All provide useful but different information.

if an os can not use GPLv3 software because apple wana hurt all opensource software by patent law then they just fuckt up!

means macos is just bad

[V!NCENT]

if an os cn not use GPLv3 software because apple wana hurt all opensource software by patent law then they just fuckt up!

means macos is just bad

+5 totally agreed. In the Linux ecosystem you need to have the software leveled due to fragmentation, but there is only one Mac OS so I'd say let it bleed.

[dargllun]

Stupid question: how are OpenGL tests carried out on OS X in those benchmarks - do they access the native OpenGL of the OS or is the X server involved? As far as I know the Mac OS X server is only some kind of add on, and wouldn't really represent the native performance. My point then would be, doesn't this benchmark compare apples and oranges?