Python Matrix Multiplication

Im trying to create a regular matrix multiplication between two huge matrices (10 * 25,000,000). My memory runs out when I do this. How can I use numpy memmap to be able to handle this? Is that even a good idea? I am not very worried about the speed of the operation, I just want to get the result, even if it means that you expect some time. Thank you for your help!

8 gbs ram, I7-2617M 1.5 1.5 ghz, Windows7 64 bit. I use the 64-bit version of everything: python (2.7), numpy, scipy.

Edit1:

Maybe h5py is the best option?

+3
source share
2 answers

np.memmap 10x10 .

, np.sum(row1 * col1).

+2

numpy.memmap numexpr! chache xD. , fortran. : python - memmaps CPU. , - , , , , pandas.hdf5 9x. , data.tofile memmap, , memmap pandas.hd5f, memmap. hdf5, - , - . , memmap 1row numpy , numpy memmap 1row .

+1

All Articles