Date: Sat, 16 Oct 1999 21:26:31 -0600 (MDT) From: "Kenneth D. Merry" <ken@kdm.org> To: groudier@club-internet.fr (Gerard Roudier) Cc: cdf.lists@fxp.org (Chris D. Faulhaber), gallatin@cs.duke.edu (Andrew Gallatin), scsi@FreeBSD.ORG Subject: Re: FreeBSD 3.2 / Slow SCSI Dell PowerEdge 4300 Message-ID: <199910170326.VAA64434@panzer.kdm.org> In-Reply-To: <Pine.LNX.3.95.991016133402.343A-100000@localhost> from Gerard Roudier at "Oct 16, 1999 01:54:18 pm"
next in thread | previous in thread | raw e-mail | index | archive | help
Gerard Roudier wrote... > > I have run some scenarii using multi-threaded IO streams under FreeBSD and > result is that the overall performance results have been significantly > better with tagged command queueing disabled. I used a Cheetah LVD and a > DRVS LVD. > > What seems to happen when tagged command queuing is disabled, is that the > IO scheduling policy can stay a long time (seconds) on one IO stream and > then switch to another one, and so on ... This seem to be due (just > guessing) to a side effect of the disksort algorithm that just makes the > supposed multi-threaded IO streams become a succession of single-threaded > IO streams that may each last seconds. Yes, there could be an odd interaction between running multiple streams with tagged queueing and the sorting algorithm. I can certainly see how this could present a problem. > If I am right (fill free to fix me if needed), it is actually the OS that > might be inadequate for tagged command queuing benchmarking. Bug or > feature ? ;-) If you'd like to test your theory about the disksort algorithm, try replacing the bufqdisksort() in dastrategy() (sys/cam/scsi/scsi_da.c) with bufq_insert_tail(). That way, the read/write requests will go through in FIFO order. Ken -- Kenneth Merry ken@kdm.org To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199910170326.VAA64434>