Date: Tue, 4 Apr 95 18:45:56 MDT From: terry@cs.weber.edu (Terry Lambert) To: rgrimes@gndrsh.aac.dev.com (Rodney W. Grimes) Cc: matt@lkg.dec.com, vernick@cs.sunysb.edu, freebsd-hackers@FreeBSD.org Subject: Re: PCI/EISA/ISA performance Message-ID: <9504050045.AA24190@cs.weber.edu> In-Reply-To: <199504042235.PAA08721@gndrsh.aac.dev.com> from "Rodney W. Grimes" at Apr 4, 95 03:35:21 pm
next in thread | previous in thread | raw e-mail | index | archive | help
> > Whose thumb are you looking at? 8-). > > Yours has a few facts wrong :-(. You're right; I looked at the EISA spec. It's different than my board settings, but then again, I may have lucked out by buying noting but high end components. My personal EISA system is running ~2 times spec (16.6MHz (50MHz/3)); it's definitely an over clocked exception. You should substitute MCA for EISA in the original post, and my numbers become happy again, since that's what I was thinking about; I just got a bunch of docs for PC busses and was reading them last night. > ISA does not have a specified clock frequency, I have seen it running > as fast as 16Mhz. Most boards die above 10Mhz, but some of the more > specialized industrial applications boards are spec'd upto 12 or 16Mhz. I've seen a lot of boards, especially serial boards and cheap floppy controllers die at anything other than 8. Many older floppy controllers used dividers instead of real clocks to get frequencies, with the result that you had to turn off "turbo" mode (drop from 16/12 to 8Mhz on the machine) to read commercially manufactured floppies. Actually, this is the gist of a reply I made on "questions" recently. > You'll never crank an ISA bus upto anywhere near > the speed of any of the others. Hard top speed limits are more like: > > ISA: 5MB/sec > EISA: 33MB/sec > VLB: 132MB/sec > PCI: 132MB/sec I think the VLB number here is a marketing lie. It can't be sustained at anywhere near that level because of it stealing refresh cycles; the good thing about PCI is that it handles memory contention better and gurantees cache writeback (VLB guarantess this only for master slots, and doesn't guarantee master slots). VLB is an OK video bus, but fairly crummy for anything else. VLB should come up better (40MHz) than PCI (33MHz), but this would be "marketing better" and not "technical better", IMO. I know only a few VLB cards that can actually go faster than 33. > And it would have probably done very well had IBM not required all > the stupid licenseing to use it. Buy a Tandy? 8-). Terry Lambert terry@cs.weber.edu --- Any opinions in this posting are my own and not those of my present or previous employers.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?9504050045.AA24190>