Date: Tue, 30 Apr 1996 09:34:21 -0700 (PDT) From: "Rodney W. Grimes" <rgrimes@GndRsh.aac.dev.com> To: koshy@india.hp.com (A JOSEPH KOSHY) Cc: hackers@freebsd.org, current@freebsd.org Subject: Re: lmbench IDE anomaly Message-ID: <199604301634.JAA13493@GndRsh.aac.dev.com> In-Reply-To: <199604301053.AA141971583@fakir.india.hp.com> from A JOSEPH KOSHY at "Apr 30, 96 04:23:03 pm"
next in thread | previous in thread | raw e-mail | index | archive | help
> > Running lmdd on a -current system supped around mid apr: > > ---- > >> single lmdd run on scsi disk > > krill# lmdd if=/dev/rsd0a of=internal bs=1024 count=8192 # SCSI disk, aha1542 > 8.00 MB in 11.35 secs, 721.92 KB/sec > ... > However when the same exercise is repeated with the IDE disk: > > ---- > krill# lmdd if=/dev/rwd0a of=internal bs=1024 count=8192 # IDE disk > 8.00 MB in 9.17 secs, 892.89 KB/sec ... > > Here we see a 8x degradation per process; 4x in terms of total throughput. > > Is this to be expected, or am i doing something wrong here? The system was > otherwise idle during these tests. run those command with a ``time'' in front of them as in: time lmdd if=/dev/rwd0a of=internal bs=1024 count=8192 # IDE disk You may be seeing CPU saturation effecting the ability of the system to keep the disk drive busy due to the fact that the wd class disks have to do rep insb to copy the data from the disk to memory where as the 1542 does this with bus mastering. -- Rod Grimes rgrimes@gndrsh.aac.dev.com Accurate Automation Company Reliable computers for FreeBSD
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199604301634.JAA13493>