[kwlug-disc] [Phoronix] Ubuntu 32-bit, 32-bit PAE, 64-bit Kernel Benchmarks

[Chris Irwin]

On Thu, Dec 31, 2009 at 04:45, Lori Paniak <ldpaniak at fourpisolutions.com>wrote:

> A while back, someone was asking about the performance differential between
> 32-bit and 64-bit kernels.  Here is some data for ubuntu karmic:
>
> http://www.phoronix.com/scan.php?page=article&item=ubuntu_32_pae&num=1
>
> Short version:  Not much difference unless you are serving web pages,
> doing SSL or disk I/O.  For those cases, 64-bit may be an order of
> magnitude faster.
>

There was a lot of discussion in their forum thread associated with this
article. There are all sorts of variables that people wanted expressed in
the article.

1. He was using Ubuntu, so it was i386 vs modern AMD64. So how much of this
was affected by having things like SSE, etc. There was a request to try
using an i686-targetted system (such as Arch). There was also a request to
use gentoo and actually build the system with the same capabilities for 32-
and 64-bits, but I doubt anybody wants to spend two weeks doing that ;)

2. There was an question regarding userspace. Would a 64-bit kernel booting
the stock 32-bit system provide a speed improvement? I would lean toward
'no', as I think most of the improvements were the userspace utilities being
able to take advantage of all the extra registers and capabilities of the
64-bit platform. I'm not that knowledgeable about hardware specifications,
so it's just a gut feeling.

So basically, he framed it as 32-bit vs 64-bit, which at the high level
("which iso do I download?") is correct. But below that it may be that some
of these improvements may be possible in 32-bit if said 32-bit system
targeted modern hardware capabilities, which is the more pressing question
considering the audience of that site (They're generally curious in Hows and
Whys).

Also, not to nitpick (though I do so love playing the part), but your short
summary read like "64-bit isn't worth it except in these specific
instances". I read the article as coming more across as "There are no
regressions, but you may get improved performance in specific instances".
Subtle difference, but it would affect whether somebody decided to install a
32- or 64-bit system.

Although it did make me somewhat curious. I use ecryptfs for my home
directory, and am curious about 32-vs-64 comparisons for that (in light of
the SSL test in the article).

-- 
Chris Irwin
<chris at chrisirwin.ca>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://kwlug.org/pipermail/kwlug-disc_kwlug.org/attachments/20091231/1c4055a3/attachment.htm>