Parallel matrix-vector multiplication in numpy

Asked Jan 01 '15 at 15:54

Active Jan 01 '15 at 15:54

Viewed 3,783 times

I have a (40000,40000) array A and vector v of size 40000. When i run

u = A.dot(b)

only one core is used. Is there any standard way to make it run in parallel?

Using the Anaconda distribution on RedHat. I have seen a lot of questions/answers about BLAS/PBLAS/ATLAS/OpenBLAS but i cannot find my way around it.

asked Jan 01 '15 at 15:54

Begelfor

2

http://stackoverflow.com/questions/5260068/multithreaded-blas-in-python-numpy – NPE Jan 01 '15 at 16:11
@NPE This is one of the answers i read, but it says nothing on how to install all those libraries/check if i already have them – Begelfor Jan 01 '15 at 16:29
1

Try numpy.__config__.show() – Martin Jan 02 '15 at 08:38

0 Answers0