Sciweavers


Publication

How to Increase Heap Space for Java VM in MATLAB

15 years 9 months ago
How to Increase Heap Space for Java VM in MATLAB
This is the solution of a problem I faced when I was analyzing data from a gigantic binary file. This white paper explains how to manage the memory heap for Matlab.
Added 22 Feb 2009
Updated 27 May 2009
Type White Paper
Comments (2)
ahmad_assayed.jpgMatlab memory usage
[+1]

One thing I found very useful with my 80MB bin file is to use indexing to select the portion of the signal that I am interested in (which in my case was only 26MB). use fseek & fread to read a specific slice of the data from the bin file.

hope this helps

-A

login or register to post comments
ahmad_assayed.jpgfew more nuggets
[+1]

when dealing with large amounts of data; keep in mind that the main memory challenge is to find a continous block of memory for evey matrix. Which means that using small arrays is much easier than one giant array - so slice your big array into smaller ones and do you processing on them.

- preallocate when possible
- when iterating on a matrix; iterate using the linear indexing when possible, it's way faster than using the regular row,column indexing

login or register to post comments