Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10 5/3/83; site gandalf.cs.cmu.edu
Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!think!harvard!seismo!rochester!pt.cs.cmu.edu!gandalf.cs.cmu.edu!mct
From: mct@gandalf.cs.cmu.edu (Mark Tucker)
Newsgroups: net.lang.lisp
Subject: Scientific Programming and Immediate Values
Message-ID: <289@gandalf.cs.cmu.edu>
Date: Tue, 22-Oct-85 09:52:54 EDT
Article-I.D.: gandalf.289
Posted: Tue Oct 22 09:52:54 1985
Date-Received: Thu, 24-Oct-85 08:20:31 EDT
Organization: Carnegie-Mellon University, CS/RI
Lines: 11

Most lisps implement integers as immedate data.
Roughly, a Lisp value is a 32bit quantity. The top 8 bits are a tag
describing how to interpret the low 24.  If the tag says CONS, then the low
24 point to a 64 bit hunk of storage that is a cons cell.  If the tag says
SMALL-INTEGER, then the low 24 bits @u<are the integer itself.>

Incrementing such a small integer just twiddles the bits of the 32-bit Lisp
value that you've got in your hand.  It doesn't mess with the heap unless
the SMALL-INTEGER overflows into a BIGNUM.

-- Mark Tucker