Ubuntu 12.10: 32-bit vs. 64-bit Linux Performance

[chithanh]

What I would be looking forward to, would be to have more dektop related tests like Firefox page parsing & rendering (and no, not again some meaningless javascript benchmark) or OpenOffice stuff.

What I found was a JavaScript comparison: http://www.raymond.cc/blog/mozilla-f...ared-to-32bit/
Generally, the 32 bit and 64 bit numbers are very close, with a slight edge for the 32 bit variant. Which is kind of expected given the experience with 64 bit Java & co.

Page rendering speed on 64 bit might benefit from SSE instructions in pixman, but a quick Google search turned up no meaningful comparison here. You'll have to benchmark that yourself.

[Linuxhippy]

Processors are natively 64-bit, so they must handle 64-bit address natively too. There is no reason that 64-bit address are slower to access. I even would say the contrary...

Problem is data structures grow, which means a smaller working set fits into caches and memory traffic increases.

Page rendering speed on 64 bit might benefit from SSE instructions in pixman, but a quick Google search turned up no meaningful comparison here. You'll have to benchmark that yourself.

Pixman has hand-written SSE instincs on x86 too, so this is not limited to 64-bit code.

and please - whoever mentioned x32 - forget it. As fast as you can. It was created by Intel to make Atom suck a bit less. Breaking all and everything just to make a crap CPU look better is NOT a good thing to

This has nothing to do with "make Atom suck less" - its just a matter of efficiency. X86 is an ugly instruction set with way too few registers, and x86_64 fixes that to some extent.
However x86_64 has the burden of 64-bit pointers, quite useless for userland in mobile devices (as you can still compile the kernel as 64-bit, and run 32-bit code) - which just makes software run slower (compared to x32) and means higher power consumption because of increased memory traffic.
Usually on phones and tablets, binary compatibility is not an issue, Intel often stated this is not intended for desktops.


There is usually a tendency that computation-bound code benefits of the extra and larger registers on 64-bit, while "heavy" bussiness-logic code like large java applications, firefox and so on suffer, because they typically use a lot of complex data structures, with many levels of indirection.

Java is a good example - its a very pointer-heavy language. Because 64bit pointers hurt Java apps typically a lot, modern java runtimes run on 64-bit machines with a hybrid between 32 and 64 bit mode, where pointers are stored in 32-bit and are expanded right before use - if memory conditions allow it.

- Clemens

[kokoko3k]

reeeeaaallly?

and you have some sources for that claim? because I read it here first....


and please - whoever mentioned x32 - forget it. As fast as you can. It was created by Intel to make Atom suck a bit less. Breaking all and everything just to make a crap CPU look better is NOT a good thing to do.

...and you have some sources for that claim?
If you want to deal with the large memory footprint of a native 64bit application you're certanly welcome.
But i'd rather prefer X32 abi and my cpu cache to be used better, instead.

--

Said that, Not only the benchmark compared 32 vs 64bit here, but even two different instruction sets.
A more fair comparision would be the same application compiled for the amd64 instruction set running on a 64bit and a 32bit kernel.

[rgloor]

Reading up on the X32 ABI gives you another data point -- X32 exposes the AMD64 architectural extensions (eg more registers for the compiler to use) while staying with 32-bit pointers in most cases. If you don't need an address space that requires >32-bit pointers it gives you "best of both worlds".

I'm glad you mentioned it.
Else I would have jumped in to put it on the table. ;-)

[rgloor]

@Michael

Hi Michael
The x32 ABI has been in the Kernel for a while and several compiles have added x32 capability as well.
I just wonder if there is somewhere a x32 version of a linux distro (test/alpha/beta)?

Because that would give a more appropriate comparison.

[rgloor]

...and please - whoever mentioned x32 - forget it. As fast as you can. It was created by Intel to make Atom suck a bit less. Breaking all and everything just to make a crap CPU look better is NOT a good thing to do.

I followed the x32 discussions for quite a while.
While there are some heavy lifters, which can utilize the 64bit width: There were several reports showing, that many applications don't really benefit from the larger registers, in contrary.
However, most application benefits from the enhanced instruction set. Which finally lead to the x32 ABI.

I want the most power out of my hardware.
If I can rarely benefit from 64bit width, I would prefer a better solution. That might be the x32 ABI.

However, before I am able to make a real bold statement, I would like to see some results of optimized* x32 versions.

*that might not be the case with the first interation available!

[AJSB]

I was talking about games...in special binary only...i had several problems with 64bit UBUNTU and derivatives...lib32 worked OK with some games and were a fail with some others....i never actually experimented with SLACKWARE...

[rgloor]

Problem is data structures grow, which means a smaller working set fits into caches and memory traffic increases.


Pixman has hand-written SSE instincs on x86 too, so this is not limited to 64-bit code.


This has nothing to do with "make Atom suck less" - its just a matter of efficiency. X86 is an ugly instruction set with way too few registers, and x86_64 fixes that to some extent.
However x86_64 has the burden of 64-bit pointers, quite useless for userland in mobile devices (as you can still compile the kernel as 64-bit, and run 32-bit code) - which just makes software run slower (compared to x32) and means higher power consumption because of increased memory traffic.
Usually on phones and tablets, binary compatibility is not an issue, Intel often stated this is not intended for desktops.


There is usually a tendency that computation-bound code benefits of the extra and larger registers on 64-bit, while "heavy" bussiness-logic code like large java applications, firefox and so on suffer, because they typically use a lot of complex data structures, with many levels of indirection.

Java is a good example - its a very pointer-heavy language. Because 64bit pointers hurt Java apps typically a lot, modern java runtimes run on 64-bit machines with a hybrid between 32 and 64 bit mode, where pointers are stored in 32-bit and are expanded right before use - if memory conditions allow it.

- Clemens

I agree!

So, what we really need, is some x32 systems and apps to run and compare!

[kokoko3k]

I was talking about games...in special binary only...i had several problems with 64bit UBUNTU and derivatives...lib32 worked OK with some games and were a fail with some others....i never actually experimented with SLACKWARE...

That is not an issue with 64bit itself, but with the game package and/or wrong libraries overrides and so on...

[chithanh]

@rgloor
Yes, there is Gentoo x32, Phoronix even wrote about it here: http://www.phoronix.com/scan.php?pag...tem&px=MTExNTE

But most people here have given up on waiting for Gentoo benchmarks from Phoronix, so it is unlikely that we will see any x32 benchmarks before a point&click install distro comes with x32 support.