NetNews Usenet Archive 1992 #18/NN_1992_18.iso/spool/comp/lang/cplus/12722

Internet Message Format | 1992-08-21 | 1.6 KB

Path: sparky!uunet!usc!elroy.jpl.nasa.gov!ames!agate!dog.ee.lbl.gov!network.ucsd.edu!sdcc12!sdcc3!acanter
From: [e-mail] (Adriaan Canter)
Newsgroups: comp.lang.c++
Subject: Turbo C++ Memory Managment On the PC
Keywords: Memory Heap
Message-ID: <[e-mail]>
Date: 21 Aug 92 13:22:12 GMT
Sender: [e-mail]
Distribution: ca
Lines: 23
Nntp-Posting-Host: sdcc3.ucsd.edu


Hello, I've noticed that Borland Turbo C++ (v 1.0) will actually reserve 16
bytes when the "new" function is used to allocate space for a 2 byte
integer, i.e. the following code:

int * intptr;
intptr = new int;

will allocate 16 bytes, as measured using the "coreleft()" function,
under the small memory model. A subsequent second allocation of this type will
fit the second integer into the first 16 byte space allocated, but a third call
will block off an additional 16 bytes, and so on, so that 4 two-byte integers
will take up 32 bytes of space, for example. I guess this happens because some
space is reserved to point to the next location in the heap, although the
amount seems excessive. This is pretty wasteful of course, if I want to
dynamically allocate alot of structures. Is there anyway I can control how
much memory is actually reserved when the "new" function is called? Do the
newer versions of C++ have some memory managment classes (as I've over-heard)
that can fix this? Or do I have to finagle my own scheme, overloading the
"new" and "delete" operators. Forgive me if this is a dumb question, but my
manuals don't say very much about it. Thank you.

                                        Adriaan Canter