Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!usc!alhena.usc.edu!ajayshah
From: ajayshah@alhena.usc.edu (Ajay Shah)
Newsgroups: comp.lang.c
Subject: Efficient dynamic storage question
Message-ID: <32121@usc>
Date: 20 Apr 91 20:17:23 GMT
Sender: news@usc
Organization: University of Southern California, Los Angeles, CA
Lines: 24
Nntp-Posting-Host: alhena.usc.edu
Originator: ajayshah@alhena.usc.edu


I have a program in which a matrix is allocated dynamically.
Every time I see an observation (which is a row in the matrix), I
realloc the matrix (which is float **) to contain one more row
(one more float *).  Then I malloc the row and fill up it's
elements.

These realloc calls are not interspersed with any *other* dynamic
memory activity.  I start building this matrix and
single-mindedly read observations till it takes it's final
shape.  So the program profile is 
realloc-malloc-realloc-malloc-realloc-malloc until all
observations are read in.

Question: is it much more efficient to realloc in groups of (say)
100 observations at a time instead of calling realloc once for each
observation?  Realloc might well have to copy a lot of elements in the
process of doing this, right?  This could be expensive.

-- 
_______________________________________________________________________________
Ajay Shah, (213)734-3930, ajayshah@usc.edu
                             The more things change, the more they stay insane.
_______________________________________________________________________________