Newsgroups: news.software.b
Path: utzoo!henry
From: henry@utzoo.uucp (Henry Spencer)
Subject: Re: Enhanced dbz for SunOS 4.0
Message-ID: <1989Jun30.162420.24518@utzoo.uucp>
Organization: U of Toronto Zoology
References: <17059@elroy.Jpl.Nasa.Gov> <1989Jun28.180430.2743@utzoo.uucp> <17099@elroy.Jpl.Nasa.Gov>
Date: Fri, 30 Jun 89 16:24:20 GMT

In article <17099@elroy.Jpl.Nasa.Gov> stevo@elroy.Jpl.Nasa.Gov (Steve Groom) writes:
>Keeping the whole file in memory can be better than seeking around the
>file and doing selective reads, but you have to have the memory for
>it.  Part of the beauty of mmap'ing is that you don't need lots of
>physical memory for it to be a win...

In principle the same is true of keeping the file in memory, as witness
demand paging.  Either way, only the active parts actually have to be in
memory.  In practice mmap makes it easier for dumb operating systems to
do things right, but it's a lot less portable.

Note that Geoff and I are old fogies who consider one megabyte to be one
@#$%^& of a lot of memory, and are thoroughly disgusted at the inability
of Suns to run decent multi-user timesharing with four whole meg of memory.
(When utzoo was a pdp11/44, it did all right -- not great, but all right --
with 2MB.)  When we say "plenty of memory", we don't mean gigabytes.
-- 
NASA is to spaceflight as the  |     Henry Spencer at U of Toronto Zoology
US government is to freedom.   | uunet!attcan!utzoo!henry henry@zoo.toronto.edu