How so? As far as Linux have proved all these years, it's rather fast.
Originally Posted by L33F3R
according to a colleague of mine, RC of FreeBSD have some "run-time lock diagnostic system" on by default, which would supposedly slow down the system. Has the phoronix team turned such an option off before testing?
Phoronix doesn't test the possible maximum speeds when comparing two OSes, but the speeds while in their default configurations. Because of that he didn't change to the writeback option of Ubuntu's filesystem as well.
Originally Posted by urandom
It will be nice to see the final releases comparison.
Originally Posted by Apopas
Loss of time
This benchmark is a complete loss of time, at least from the FreeBSD perspective. Benchmarking a BETA or RC version of FreeBSD is at best stupid for a very simple reason:
NOTE TO PEOPLE WHO THINK THAT FreeBSD 8.x IS SLOW:
FreeBSD 8.x has many debugging features turned on, in both the kernel and userland. These features attempt to detect incorrect use of system primitives, and encourage loud failure through extra sanity checking and fail stop semantics. They also substantially impact system performance. If you want to do performance measurement, benchmarking, and optimization, you'll want to turn them off. This includes various WITNESS- related kernel options, INVARIANTS, malloc debugging flags in userland, and various verbose features in the kernel. Many developers choose to disable these features on build machines to maximize performance. (To disable malloc debugging, run ln -s aj /etc/malloc.conf.)
This is from a file named UPDATING, located in /usr/src. If you want to test FreeBSD you should wait for the RELEASE, not RELEASE CANDIDATE.
You only proved that you are ignorant vis-a-vis FreeBSD and managed to mislead readers with the results.
This so called benchmark is useless.
The BSD release has all sorts of debugging turned on. It significantly reduces performance.
Thus, this benchmark is invalid. Doesn't show anything even remotely related to reality.
Do it again, and do it right.
Well, Phoronix is a Linux centric site. What did you expect?
Originally Posted by clau
Also this is a benchmark - not a test nor a real life experience.
Though I prefer Linux (mainly due to GNU user land), I have a deep respect for the stability of *BSD.
Also, on the occasions when I had to work with FreeBSD, I personally never experience any huge performance problems on FreeBSD. It is mostly slower than Linux, but I pretty much never seen a case when FreeBSD would DoS itself. Linux in past did that to itself quite often: focus of development is way too often on performance rather than correctness and stability. (That's why too programs under Linux never terminate with "memory allocation error" - they are simply killed by kernel. It's faster to kill application asking for more memory than to return properly an error code to it.)
Disk I/O is a classical example of BSD v. Linux difference of stability v. performance: BSD attempts to always write meta-data before writing file data, while Linux would happily keep dirty meta-data in cache. It's faster that way, but you can never be sure what happens to your files on crash. In BSD, when program opened, wrote and closed file, after following immediately power outage file would be highly likely intact. In Linux - YMMV. Most of the time file would be missing. (N.B. This is a different issue from recent Ext4 fiasco.) That's why in Linux I/O is so fast - but you better have a UPS under your table.
I still do not like BSD user land, but bashing BSD performance IMO is simply stupid.
I just hope this finally shuts up the 'freebsd is faster than linux' crowd. They told that lie until Felix von Leitner tested the BSDs againat 2.6.0 - and they got ass raped. Whine, whine, not fair they complained. True - since Felix did everything in their favour.
But shortly afterwards the idiotic stuff started again....
no matter that almost all tests show that linux is superior....
Was FreeBSD debuging disabled?
Pre-releases of FreeBSD ship with lots of compiled in debugging code that negatively impacts performance. If it was not disabled, the benchmark is