MathGroup Archive 1995

[Date Index] [Thread Index] [Author Index]

Search the Archive

question on caching

  • Subject: [mg2700] question on caching
  • From: loh at (Vincent Kwok-Shiang Loh)
  • Date: Sat, 9 Dec 1995 01:52:16 -0500
  • Approved:
  • Distribution: local
  • Newsgroups: wri.mathgroup
  • Organization: MIT Microsystems Technology Laboratories

Hi all,
  I saw a thread on caching and it led me to think about my code:

I built a general tree structure using caching to implement different
nodes on the tree:

	node[x, y] returns the contents of node(x, y)

This seemed like a fast and easy way to build tree structures.
However, I've been aware of memory constraints.  Realistically, how
much amount of caching can I do before I have to start worrying?

my machine setup is:
	486 NT 3.51
	52Mb ram ( :-) )
	~196Mb swap device.

I don't think I am hitting memory problems now, because my disk is not
grinding to death, and my memory load indicator seems to be doing fine.

Anyway, I would like to know how efficient in terms of storage and
speed the cache is.

Also, I had originally built a tree using a list structure.  I was
getting concerned with carrying a huge complicated list around.  since
mma works with expressions, I was afraid that my variable was being
evaluated constantly and many duplicates of large, large trees were
being created and destroyed when functions were called.  I think that
this is in fact not the case, but I'm not sure how mma does this.  

If anybody could fill me in on any of the issues, I would appreciate


  • Prev by Date: Re: Scientific Web - Speedcomparison Mathematica 2
  • Next by Date: Box counting dimension for fractals via Mathematica??
  • Previous by thread: Re: Comparison of MMA on Various Machines
  • Next by thread: question on caching