Are you using the same part of the disk for both tests? - Yes on both, is an old scsi controller but supported ( I checked the HLC )
Is the OpenBSD fs using softdep? - How can i check this? What is the amount of memory in the machine? - 2Gb How many runs is this the average of? - On linux in the same conditions ( clean install, dd from the same partition, etc.. ) i get 17MB/s Is the input_file created immediately before the test? Yes on both Is the machine running other processes at the same time? Yes, I made a new instalation with minimum but i don't disable anything. I want to test it on daily conditions to see the real performance. Thanks. 2007/7/18, francisco roque <[EMAIL PROTECTED]>: > > You should include the details of your test, such as: > > Are you using the same part of the disk for both tests? > Is the OpenBSD fs using softdep? > What is the amount of memory in the machine? > How many runs is this the average of? > Is the input_file created immediately before the test? > Is the machine running other processes at the same time? > > All of those factors (and surely more i forgot) can influence the results. > Personally i prefer bonnie++ and IOzone to check general disk performance. > > Good luck, > > -f > http://www.blackant.net/ > > On Wed, 18 Jul 2007, Tang Tse wrote: > > > retaking this thread, > > > > I got the same issue, very poor disk performance comparing openbsd 4.1with > > linux 2.6.22 > > > > # time dd if=input_file of=file_out bs=1024 count=1024000 > > input_file is 1GB > > > > On OpenBSD box it takes 4min, transfering about 3,3MB/s > > On linux 2.6.22 it takes 1min, transfering about 17MB/s > > > > The difference is very very big.. > > > > > > 2007/7/3, Martin Toft <[EMAIL PROTECTED]>: > >> > >> On Tue, Jul 03, 2007 at 10:20:18PM +1000, Jonathan Gray wrote: > >>> On Tue, Jul 03, 2007 at 01:49:09PM +0200, Martin Toft wrote: > >>>> Disk I/O is the only test where I use different programs (hdparm and > >>>> dd), as I couldn't find a port/package of hdparm for OpenBSD. > >>>> Still, I think the results are so different that they set off "alarm > >>>> bells" -- 8.5-8.7 MB/s vs. 45-46 MB/s. > >>> > >>> Well at least use dd in both cases and use the same kinds of buffered > >>> or unbuffered devices/files. > >>> > >>> I imagine the results will be diferrent if you dd from a file to > >>> /dev/null for example. > >> > >> You're absolutely right. On OpenBSD, dd'ing a file actually gives an > OK > >> result: > >> > >> $ dd if=KNOPPIX_V5.0.1CD-2006-09-25-DA.iso of=/dev/null > >> 1433280+0 records in > >> 1433280+0 records out > >> 733839360 bytes transferred in 22.626 secs (32432248 bytes/sec) > >> > >> 30.93 MB/s that is. As I can't figure out how to mount my OpenBSD > >> partitions on KNOPPIX, I can't do the same test in that environment. > >> Thanks for pointing out that the previous comparison was unfair. > >> > >> It seems that I can't really be disappointed with my OpenBSD disk I/O > >> now, only the system's number crunching abilities. I would like to > >> remind you, that I could squeeze a lot more CPU power out of the laptop > >> with OpenBSD -current about a month ago, so in some way, I suspect that > >> some crucial code has been changed in the meantime. > >> > >> Martin
