Re: 3.0 = Rip Roaring Resource Hog :-(
- To: mathgroup at smc.vnet.net
- Subject: [mg5129] Re: [mg5094] 3.0 = Rip Roaring Resource Hog :-(
- From: jpk at apex.mpe.FTA-Berlin.de (Jens-Peer Kuska)
- Date: Wed, 6 Nov 1996 01:32:47 -0500
- Sender: owner-wri-mathgroup at wolfram.com
> It is starting to sink into my brain that Mathematica is extremely wasteful > of system resources. For the first time, I am thinking about moving my > work over to MATLAB. I need a program that knows what machine numbers are! schnipp, schnipp > MATHEMATICA 3.0 TAKES 9 MEGABYTES TO STORE 0.5 MEGABYTES OF DATA. This is > an inefficiency ratio of 18:1. I AM HAVING SERIOUS DOUBTS ABOUT THE > ABILITY OF THIS PROGRAM TO DEAL WITH REAL LABORATORY DATA SETS. Hi Mark, Mma has the most general descripton of expressions. This general structure had it's price. Mutch of the beauty of Mma and it's programming language relies on this general structure. You must keep in mind that Mathematica is a Computer-Algebra System and *not* - a data analysis system - graphics system Yes, You can use it for data analysis and graphics and You can eat ice cream with a knife. You can`t have the powerful general programming language *and* the most compact data representation for Your private kind of data. And so, please WRI keep all your efforts on the general data structure and the powerful programming language -- dont mind about memory (its cheep and will become cheeper every day). Jens PS: I use Mma over years to process my data from particle/plasma simulations. Every run of my simulation program produces 300 Mbyte of binary data -- particle coordinates snapshoots, 3D magnetic & electric fields. All works fast and perfectly since I can label my Mma graphics with allmost all symbols I can imagine. But Mma had never seen my data as lists, list of lists or what ever. I had a MathLink programm that reads and keeps my data. The C-program has it's own data structure to keep some megabytes in a compact way. Typical I have 3d array of the (electron) density. The data are stored in a C-array in the MathLink client. The interface to the data is a interpolation of the data at every space point. I can access my data via a MathLink function Interpolate3D[x,y,z] and I had never problems with that method. You see -- that is the price for my compact data I have to program in C, deal with pointers and memory allocation and I hardly miss Mma /@, @@, _?NumberQ.