Am 11/20/12 22:56, schrieb Miroslav Lachman: > Mike Jakubik wrote: >> On 2012-10-12 05:21 PM, Adrian Chadd wrote: >>> On 12 October 2012 11:10, Miroslav Lachman <000.f...@quip.cz> wrote: >>>> I don't like comparing Release Candidates without any details about >>>> config, >>>> but the fact that DF 3.2 is much better than DF 3.0 is interesting. >>>> And they >>>> are very close to performance of Scientific Linux 6.2. >>> >>> Hey cool! And FreeBSD-9.1 is on there and doing worse than Linux and >>> Dragonfly BSD. I wonder why that is. >>> >>> Lemme cross post this a little to see what people think. >>> >>>> >>>> http://lists.dragonflybsd.org/pipermail/users/2012-October/017536.html >>>> >>>> Graphs are available as PDF attachments >>>> >>>> >>>> http://lists.dragonflybsd.org/pipermail/users/attachments/20121010/7996ff88/attachment-0002.pdf >>>> >>>> >>>> >>>> http://lists.dragonflybsd.org/pipermail/users/attachments/20121010/7996ff88/attachment-0003.pdf >>>> >>>> >> >> These numbers show very significant improvements. Any >> possibility/interests in porting this scheduler to FreeBSD or this too >> much work? I know many have and still complain about our current >> scheduler. > > Just for the record. There is Phoronix benchmark comparing DF 3.0.3, > 3.2.1 and Ubuntu 12.10 showing that scheduler tweaks was good for > pgbench, but not so significant for other tasks. > > http://www.phoronix.com/scan.php?page=article&item=dragonfly_linux_32 > > Miroslav Lachman
It is always this pgbench benchmark that gives the numbers and the scientific area seems to get a shortcut all the way. Well, by zapping through Phoronix' benchmarks I found no recent benchmark comparing FreeBSD 9.1-PRE, DargonFly 3.0 and 3.2 as well as Ubuntu 12.04/10 or Scientific Linux 6.2, only older ones: Linux Red Hat 6.2 versus Ubuntu 12.04 LTS: http://www.phoronix.com/scan.php?page=article&item=rhel_ubuntu_62&num=1 Mac OS X vs. Ubuntu 12.04 LTS: http://www.phoronix.com/scan.php?page=article&item=macosx_108dp1_ubuntu&num=1 Oracle Linux 6.1 versus FBSD 9.0: http://www.phoronix.com/scan.php?page=news_item&px=MTAyNzA I find it hard to make those benchmarks comparable, since the test parcours seems to be inconsistent from benchmark to benchmark. The HIMENO bench occurs late, CRAY bench is not available in most benchmark tests and even the NAS test isn't very often present. It is not easy to find a clear answer for questions like "what compiler and compiler options has been used ... et cetera. I'm impressed by the DragonFly approach and performance jump and I'm disappointed having FreeBSD again far behind Linux - but it is the PostgreSQL benchmark. It seems that in most cases, Ubuntu Linux is always ahead of every other OS at the moment - if someone "believes" the Phoronix benchmark results. O. Hartmann
signature.asc
Description: OpenPGP digital signature
