[Numpy-discussion] Where is the instruction of installing numpy with Intel lib MKL?
Thu Mar 26 14:32:10 CDT 2009
BTW, this timing on a core 2 Duo 2.0GH laptop ,with the Enthought Python
Distribution, is around 0.2 second.
Ben Park wrote:
> I have spent many hours trying to do this, to no avail. The numpy
> installation I got didn't seem to link to the MKL library. A 1000x1000
> matrix multiplication took 8 seconds.
> import numpy as N
> import numpy.random as RN
> import time
> m = 1000
> #m = 2000
> #m = 10000
> #m = 30000
> #m = 40000
> X = N.zeros((m,m),'f')
> for i in range(m):
> X[i,:] = RN.random((m,)).astype('f')
> #X = RN.random((m, m)).astype('f')
> t1 = time.time()
> Y = N.dot(X, X)
> t2 = time.time()
> print 'dt = %f' %(t2-t1)
> The result is about 8 seconds on a Intel Core 2 Duo 2.0GH laptop.
> Certainly not fast.
View this message in context: http://www.nabble.com/Where-is-the-instruction-of-installing-numpy-with-Intel-lib-MKL--tp22728276p22729404.html
Sent from the Numpy-discussion mailing list archive at Nabble.com.
More information about the Numpy-discussion