From owner-freebsd-hackers Mon Dec 7 13:14:14 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id NAA16795 for freebsd-hackers-outgoing; Mon, 7 Dec 1998 13:14:14 -0800 (PST) (envelope-from owner-freebsd-hackers@FreeBSD.ORG) Received: from terra.Sarnoff.COM (terra.sarnoff.com [130.33.11.203]) by hub.freebsd.org (8.8.8/8.8.8) with SMTP id NAA16788 for ; Mon, 7 Dec 1998 13:14:12 -0800 (PST) (envelope-from rminnich@Sarnoff.COM) Received: (from rminnich@localhost) by terra.Sarnoff.COM (8.6.12/8.6.12) id QAA09577; Mon, 7 Dec 1998 16:14:08 -0500 Date: Mon, 7 Dec 1998 16:14:08 -0500 (EST) From: "Ron G. Minnich" X-Sender: rminnich@terra To: hackers@FreeBSD.ORG Subject: Re: write combining memory on pci In-Reply-To: <366C43DD.A7DA16F4@tdx.co.uk> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG On Mon, 7 Dec 1998, Karl Pielorz wrote: > Both my P-Pro systems have this option, and AFAIK it worked as above (i.e. > 'all on' or 'all off')... Win NT survives with it on, until I try to use the > sound card, and FreeBSD, Linux & Win'95 all die within seconds with it on Hmm, that is bad news. I have some cards from a company that only run fast with write combining turned ON. However, if they crash the OS with it on, it might not be such a good deal. Guess I'll have to use the cards as a master, not a target for PIO writes. As ever, the rule with PCI: don't get the CPU involved in data transfer :-) > Which is a bit of a pitty, as I also seem to remember reading it gives a huge > performance win on writes ;-) - Maybe there is such a thing as 'too fast'... something on order of at least 2x. It would be nice to have ... ron To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message