Date: Tue, 12 Nov 1996 13:24:27 -0800 (PST) From: "Rodney W. Grimes" <rgrimes@GndRsh.aac.dev.com> To: michaelv@MindBender.serv.net (Michael L. VanLoon -- HeadCandy.com) Cc: scrappy@ki.net, smp@csn.net, smp@freebsd.org, current@freebsd.org Subject: Re: GigaByte GA-586DX-512 Motherboard Message-ID: <199611122124.NAA17321@GndRsh.aac.dev.com> In-Reply-To: <199611121756.JAA06618@MindBender.serv.net> from "Michael L. VanLoon -- HeadCandy.com" at "Nov 12, 96 09:56:09 am"
next in thread | previous in thread | raw e-mail | index | archive | help
> > >>I recommend 64meg of parity. Don't really have direct knowledge of EDO vs not. > > > Hrmmm...I had thought that EDO RAM was better because it had > >'on chip' cache or something like that ... > > Uh, no. EDO RAM simply takes less clock cycles per access than > standard (or FPM) RAM. You should go read some technical specifications before you make such statements. The actual access time changes very very little between EDO (``Extended Data Out'') and FPM (``Fast Page Mode''). The only real effective difference is that EDO memory holds it's data outputs valid after CAS goes back high, allowing a wider and slopper data latching. This _can_ lead to elimination of wait states in poor designs, but does zilch in good designs, since good designs don't need the added 15nS of valid data out as they latch the data in the 5 to 6nS window you get when running FPM at minimum timings. Note that the cycle time for EDO vs FPM is IDENTICAL for every manufactures data sheet I have looked at. Real memory bandwidth is a direct function of cycle time, and has nothing to do with access times. Access times are for wimps, real memory designers work with cycle times :-) :-). You can not transfer 1MB of data out of EDO memory any faster than you can transfer 1MB of data out of FPM. Period. Look over the data sheets, study memory design, _know_ before you speak. Also note that the ``access'' time difference is usually less than 5nS and due to other timing constraints you can rarely if ever take advantage of this. -- Rod Grimes rgrimes@gndrsh.aac.dev.com Accurate Automation, Inc. Reliable computers for FreeBSD
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199611122124.NAA17321>