Date: Thu, 13 Jun 1996 14:26:18 -0500 From: steve farrell <spfarrel@midway.uchicago.edu> To: "stable@freebsd.org" <stable@FreeBSD.org> Subject: Re: CTM back in sync. Message-ID: <199606131926.OAA00293@meno.uchicago.edu> In-Reply-To: Your message of "13 Jun 1996 09:36:06 CDT." <n1377463920.99467@Richard Wackerbarth>
next in thread | previous in thread | raw e-mail | index | archive | help
ok -- i replaced my NE2000 with a 3C509, hoping with the guess that a bad ethernet card was causing the lockups i was experiencing. i have not had the card in nearly long enough to judge if this has fixed the problem, but i did a few little benchmarks, and it tests well (650K/sec). but it's a little odder than that. if i'm ftping to an nfs mounted filesystem (on solaris 2.5/sparc), it gets about 100K/sec and makes the drive in the sparc gargle quite a bit... almost thrash. kind of odd, probably sun's fault =). what i'm posting to this list about is a far odder occurence: if i write to my EIDE disk, i get the aforementioned 650K/sec. but if i write to my SCSI disk, i get an abysmal 20K/sec. the system is a 486 w/48Meg of ram, so i wonder if this is some odd effect of having two cards attempting to do DMA through the bounce buffers? it this expected behavior??? (the SCSI disks otherwise seem to give normal performance...) (system is -stable which i supped just before recompiling for the 3Com card -- i.e., about an hour ago - from sup.freebsd.org...)
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199606131926.OAA00293>