Path: utzoo!attcan!uunet!husc6!mailrus!uflorida!novavax!proxftl!bill
From: bill@proxftl.UUCP (T. William Wells)
Newsgroups: comp.lang.c
Subject: Re: memory allocation
Message-ID: <733@proxftl.UUCP>
Date: 10 Sep 88 13:58:38 GMT
References: <1262@micomvax.UUCP>
Reply-To: bill@proxftl.UUCP (T. William Wells)
Distribution: comp.lang.c,comp.os.misc
Organization: Proximity Technology, Ft. Lauderdale
Lines: 16
Summary:
Expires:
Sender:
Followup-To:
Keywords:

In article <1262@micomvax.UUCP> belanger@micomvax.UUCP (Robert Belanger) writes:
: Hi everybody,
:
:       We are in the process of speeding up memory allocation. Because we
: have a CPU with a linear address space, but no (easy) access to disk. We have
: four(4) megabytes of DRAM available. What we want to do is avoid any system
: in order to speed up memory allocation and also reduce the overhead associated
: with memory allocation.

One scheme comes to mind for fast memory allocation: the "buddy
system allocator".  This is *extremely* fast.  The drawback is
that you must allocate in blocks of certain sizes.  For example,
the standard method uses blocks of size k*2**n, where k is a
constant and n is a parameter.  Look in Knuth or in "Data
Structures and Algorithms", Aho, Hopcroft, and Ullman, for a
description of this and other memory allocation systems.