To date I implemented a few functions I needed, i.e. some C-like operators and circshift:
inplace(x, y, '+=');
inplace(x, y, '.*=');
inplace(x, ns, 'circshift');
that do not allocate memory and compared to the matlab equivalent
x = x + y;
x = x .* y;
x = circshift(x, ns);
are (when working with arrays above 300-400 MB) much faster (tenths of second instead of a few minutes) because the pc does not start swapping to the disk.
The matlab equivalents are currently optimized (performed in-place) only in M-code, but not at the command line.
See http://blogs.mathworks.com/loren/?p=36 for a discussion about it.
IMPORTANT: this piece of code infringes the recommendation that mex files should not overwrite input data.
Luca Citi (2021). inplace_ (https://www.mathworks.com/matlabcentral/fileexchange/12626-inplace_), MATLAB Central File Exchange. Retrieved .
Luca, the cvlib_mex submission already did dat with further advantages:
1. It works with all data types (except uin32) and not only with doubles.
2. With pentiums it uses the Matlab MKS library to do matrice operations
Find the treasures in MATLAB Central and discover how the community can help you!Start Hunting!