Date: Wed, 10 Sep 1997 00:00:28 +0200 From: Mats Lofkvist <mal@algonet.se> To: tlambert@primenet.com Cc: current@FreeBSD.ORG Subject: Re: lousy disk perf. under cpu load (was IDE vs SCSI) Message-ID: <199709092200.AAA21008@kairos> In-Reply-To: <199709091834.LAA28493@usr04.primenet.com> (message from Terry Lambert on Tue, 9 Sep 1997 18:34:13 %2B0000 (GMT))
next in thread | previous in thread | raw e-mail | index | archive | help
You were right about tagged queueing not being enabled.
Unfortunately turning it on didn't seem to help.
I also tried with the 2.2.2-r generic kernel with the same results
(no tagged queueing enabled in that one).
I will now try to reassign the interrupts as Stefan suggested.
_
Mats Lofkvist
mal@algonet.se
Changes in config file to enable tagged queueuing:
# Tagged queueing on the ncr controller => no FAILSAFE
#options FAILSAFE
# Tagged queueing on the ahc
options AHC_TAGENABLE
# These two are supposed to be an improvement also
options AHC_SCBPAGING_ENABLE
options AHC_ALLOW_MEMIO
Dd test on quantum viking connected to ahc, unloaded:
garm>dd if=/dev/rsd0 of=/dev/null bs=4k count=8000
8000+0 records in
8000+0 records out
32768000 bytes transferred in 3.068474 secs (10678924 bytes/sec)
Loaded (nice -19 loop &):
garm>dd if=/dev/rsd0 of=/dev/null bs=4k count=800
800+0 records in
800+0 records out
3276800 bytes transferred in 69.099672 secs (47421 bytes/sec)
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199709092200.AAA21008>