[Numpy-discussion] Huge arrays

[Sturla Molden]

Daniel Platz skrev:
> data1 = numpy.zeros((256,2000000),dtype=int16)
> data2 = numpy.zeros((256,2000000),dtype=int16)
>
> This works for the first array data1. However, it returns with a
> memory error for array data2. I have read somewhere that there is a
> 2GB limit for numpy arrays on a 32 bit machine but shouldn't I still
> be below that? I use Windows XP Pro 32 bit with 3GB of RAM.

There is a 2 GB limit for user space on Win32, this is about 1.9 GB. You 
have other programs running as well, so this is still too much. Also 
Windows reserves 50% of RAM for itself, so you have less than 1.5 GB to 
play with.

S.M.