Date: Sun, 10 Jan 1999 23:09:39 +0100 From: Stefan Esser <se@mi.uni-koeln.de> To: lcremean@tidalwave.net Cc: Kenneth Wayne Culver <culverk@wam.umd.edu>, freebsd-current@FreeBSD.ORG, Stefan Esser <se@FreeBSD.ORG> Subject: Re: ide_pci Message-ID: <19990110230939.A43344@dialup124.mi.uni-koeln.de> In-Reply-To: <19990110024647.A50225@tidalwave.net>; from Lee Cremeans on Sun, Jan 10, 1999 at 02:46:47AM -0500 References: <Pine.GSO.3.95q.990110000744.12864D-100000@rac4.wam.umd.edu> <19990110024647.A50225@tidalwave.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On 1999-01-10 02:46 -0500, Lee Cremeans <lee@st-lcremean.tidalwave.net> wrote: > beautifully, pulling down 13+ MB/s from my Seagate Medalist Pro 9140, but > writes were kinda recalcitrant, and I think that may have been because I > overclocked the PCI bus on my board -- I need to test it at 33 MHz. I can't EIDE DMA writes are directly coupled to the external clock and will come out way too high if you overclock your CPU. Some EIDE tolerate 10% to 15% higher bus clocks, but some fail even at 5% more than specified! But it is possible to reduce the EIDE clock by choosing a higher divider value in some controller register, unless you are using an external clock of more than 100MHz ... Going to DMA Mode 1 transfers should make writes work at half the UDMA data rate times your overclocking factor, which should still be fast enough ;-) Regards, STefan To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19990110230939.A43344>