Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!ucbvax!ucsd!ucsbcsl!dougp@sbphy.ucsb.edu
From: dougp@sbphy.ucsb.edu
Newsgroups: comp.sys.amiga
Subject: Re: Hard drive & chip contention
Message-ID: <1384@hub.ucsb.edu>
Date: 28 Mar 89 03:26:06 GMT
Sender: news@hub.ucsb.edu
Organization: UC, Santa Barbara. Physics Computer Services
Lines: 15


>In article <689@boing.UUCP> dale@boing.UUCP (Dale Luck) writes:
[stuff about how bus contention is handled in fast ram deleted]

Wouldn't things be improved if you had real fast ram at 0xC00000
instead of slowfast ram like the 2000 currently has? This would
put most of the OS structures and code in real "fast" ram, then
the 68000 would be less likely to be tring to access ram which
the bus arbitrator thinks is chip ram, and less likely to get
stuck waiting for the chips to finish.

So if, when the new chips come out and the slowfast 0xC00000 ram
becomes chip ram, you buy a 1 meg 0xC00000 fast ram board, will
disk reads while displaying 640 wide 16 color screens be closer 
to normal speed?