Ubuntu 13.10 32-bit vs. 64-bit Performance

[bridgman]

I'm sorry, how is a red and green graph discriminatory?

I can think of two possibilities :

1. The poster could be talking about colour blindedness -- Red-green is the most common form in humans, and affects males more than females.

2. The poster could be a Red Green fan, but neither handsome nor handy.

[synaptix]

Also, most of network/web tests are based on ipv4 which is 32bit, so there's no way to optimize it in a 64bit system.
Phoronix, could you do more testing in ipv4 between 32/64bit systems?

How is there no way to optimize it?

[enihcam]

tcp/ip stack performance is also needed for 32/64bit comparison.

Encoding is not vastly different from all the other microbenchmarks, its more or less the same type of workload (kind of number-crunching) - where you spend a lot of time in quite small and highly optimized parts of a programm.

Serving static webpages is a completely different workload, where a lot of different code is executed with complex data-structures.
So serving web-pages is actually a lot more like rendering web-pages or loading/saving large documents in LibreOffice and other everyday-stuff.

Also, most of network/web tests are based on ipv4 which is 32bit, so there's no way to optimize it in a 64bit system.
Phoronix, could you do more testing in ipv4 between 32/64bit systems?

[Linuxhippy]

No? There is literally zero performance difference between them, and Mint is less up-to-date on software since the releases are delayed compared to Ubuntu -- and this is even more the case for the Xfce edition.

Encoding is not vastly different from all the other microbenchmarks, its more or less the same type of workload (kind of number-crunching) - where you spend a lot of time in quite small and highly optimized parts of a programm.

Serving static webpages is a completely different workload, where a lot of different code is executed with complex data-structures.
So serving web-pages is actually a lot more like rendering web-pages or loading/saving large documents in LibreOffice and other everyday-stuff.

[gens]

i like green

anyway, it dosent matter what distro for most of these tests if not all of them
(3D stuff would maybe have more fps, but same %'s for comparison)

kernel dosent matter in most and libc is the same in all

[synaptix]

red-green graf is discriminatory against a minor quantity of men.
Why not red-blue for example?

I'm sorry, how is a red and green graph discriminatory?

[ulenrich]

red-green graf is discriminatory against men

red-green graf is discriminatory against a minor quantity of men.
Why not red-blue for example?

[synaptix]

No? There is literally zero performance difference between them, and Mint is less up-to-date on software since the releases are delayed compared to Ubuntu -- and this is even more the case for the Xfce edition.

Really, is that why I can get more FPS on games running Mint 15 and I get better system responsiveness and my shutdowns don't hang like it does on Ubuntu?

Oh and is that why all my packages and software matches that of current up to date version on Ubuntu as well?

[Calinou]

Would be nice to see some Linux Mint benchmarks, granted it's still based on Ubuntu it works a LOT better than Ubuntu most times.

It's his website, he can do whatever the hell he wants so it doesn't bother me that he only benchmarks Ubuntu most of the time.

No? There is literally zero performance difference between them, and Mint is less up-to-date on software since the releases are delayed compared to Ubuntu -- and this is even more the case for the Xfce edition.

[GreatEmerald]

Mandriva(or whatever it's calling itself these days)

Mageia.

Sad that there are so many computational tests like encoding or micro-benchmarks like c-ray.

Except for the games, there aren't *any* tests which perform typical workload like bootup-time, webbrowser-performance, ....

Encoding is quite relevant even for regular people. Though you're right that static webpage serving kind of isn't.