PC-BSD 8.0 vs. Kubuntu 9.10 Benchmarks
Phoronix: PC-BSD 8.0 vs. Kubuntu 9.10 Benchmarks
PC-BSD 8.0 was released last week and while we have already delivered FreeBSD 8.0 benchmarks including against Debian GNU/kFreeBSD and Fedora / Debian / OpenBSD / OpenSolaris for which PC-BSD is based, we took this opportunity to deliver a fresh set of *BSD benchmarks. In this article we have benchmarks of PC-BSD 8.0 x64 against Kubuntu 9.10 x86_64.
So, UFS sucks that badly hey? Or is there something else that affects filesystem related tests?
they use different version of gcc, and all known, gcc 4.4.x have better optimization than 4.2.1.
most of these benchmark are focus on CPU calculation.
Could do some other benchmark. like tcp/udp network performance, mysql performance, bind/fork performance. etc.
Hey Michael, how many hard drives have died on you prematurely from constant formatting/reformatting of various file systems and OS'?
Yeah, some people are not understanding the meaning of "Out-of box" benchmarking...
Originally Posted by L33F3R
I registered just to post in this thread.
Such benchmarks would be nice to see, although they probably would need to be done with both operating systems being compiled with the same version of GCC for them to meaningfully isolate the pros and cons of each OS.
Originally Posted by Deng XueFeng
I run Gentoo Linux. Tell me more about this out-of-box benchmark of which you speak. It is a foreign concept to me.
Originally Posted by Rip-Rip
To an extent. There is a 80/20 rule to most things. 80% of the people (even Gentoo people) won't know all the knobs to turn. They rely on the original packager (or ebuild maintainer) to get most things set up from there.
Originally Posted by Shining Arcanine
Same for compilers. There are differences between icc/gcc/sun studio that all have particular strengths. For most users however, they rely on the upstream maintainers to do the right technology selection for them.
But again, it's all about what you are trying to test. Phoronix tries to test the "full system" as provided by the vendor. If you want to test the kernel/libraries as shipped between distributions, then yes - the same compiler version is probably needed. But Michael is testing the "full system", not the lower parts.
formatting a drive is not a stress test; it's not like every block has to be rewritten
Oh. I've heard of a couple people losing drives because of constant reformatting. Would that just be due to faulty/cheap drives?
Tags for this Thread