Re: Benchmark (Phoronix): FreeBSD 9.0-RC2 vs. Oracle Linux 6.1 Server

Am 16.12.2011 07:06, schrieb Alex Kuster:
> On 12/16/2011 02:41, Arnaud Lacombe wrote:
>> Hi,
>>
>> On Thu, Dec 15, 2011 at 2:32 AM, O. Hartmann
>> <ohartman_at_zedat.fu-berlin.de>  wrote:
>>> Just saw this shot benchmark on Phoronix dot com today:
>>>
>>> http://www.phoronix.com/scan.php?page=news_item&px=MTAyNzA
>>>
>> it might be worth highlighting that despite Oracle Linux 6.1 Server is
>> using a kernel + compiler almost 2 years old, it still manages to
>> out-perform the bleeding edge FreeBSD :-)

No, there was no measurement of Oracle Linux 6.1 compared to a normal
FreeBSD installation as has already pointed out by quite a number of people.

>> Now, from what I've read so far in this thread, it seems that a lot of
>> people are still in abnegation...
>>
>> my 0.2c,
>>   - Arnaud
> 
> This smells like flamebait ...
> Because everyone with a little love or knowledge about benchmarking
> would realize that the benchmark is all wrong, and not only that ...
> they say that the benchmark tests defaults and ZFS, afaik is far from
> being a default.

Yes, and a default installation of FreeBSD (with UFS2 and SU or SU+J)
would have allowed to run the *exact same* binaries used in the Linux
test by just recursively copying the Linux root to /compat/linux (and
loading linux.ko, of course). There is some emulation overhead (more
pathes are searched, for example), but FreeBSD compared well under
realistic loads in prior tests.

The problem with a number of the tests (obviously measuring the amount
of dirty buffers allowed by the kernel before a generating program is
throttled back to prevent loosing valuable buffer cache contents) does
also lead to very misleading results (since they do not measure a steady
state load situation common on a server).


We have gone through this topic a number of times (as a search for
Phoronix on the mail archives schould be able to reveal).

There may be performance advantages for either OS compared to the other,
but most of the Phoronix tests are totally unsuitable to find them, even
when performed under fair conditions (e.g. same compiler version,
comparable file system).

STefan