home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.os.mach
- Path: sparky!uunet!usc!elroy.jpl.nasa.gov!ames!pacbell.com!att!cbnewse!cbnewsd!bgbg
- From: bgbg@cbnewsd.cb.att.com (brian.g.beuning)
- Subject: Re: multi processor 386 and mach
- Organization: AT&T
- Date: Fri, 21 Aug 1992 03:04:01 GMT
- Message-ID: <1992Aug21.030401.15123@cbnewsd.cb.att.com>
- References: <mcrosbie.714068043@unix1.tcd.ie>
- Lines: 22
-
- From article <mcrosbie.714068043@unix1.tcd.ie>, by mcrosbie@unix1.tcd.ie (Mark J. Crosbie):
- >
- > Simply put, to buy a machine with more than 1 386 would be very
- > expensive as you would be paying for the multi-processor bus which would
- > have to be quite complex if caches were used on each board.
- > It would also have to be recognised by Mach, as each bus implementation
- > is different. If Mach didn't support it, you could forget it!
- >
- > In my humble opinion, it's a lot simpler to buy a FAST (50Mhz) 486
- > machine with a fast
- > SCSI disk and lots of memory to run Mach (>8 MB) than it is to buy
- > a multi-processor system.
-
- Thanks for answering my query. I guess I was showing my hardware
- ignorance by not knowing the bus could be such a price factor.
-
- I have been getting very interested in Mach on a 386 as being
- the first "multi-processing for the masses". But it sounds
- like that day is not here yet. Oh well, I'll just need to be
- content with blazing speed instead of finesse :-)
-
- Brian Beuning
-