arpack.eigen gives incorrect result

[kyleabeauchamp]

I have previously posted a similar report, I'm following up with a test case. Basically, sometimes arpack.eigen will give an incorrect result.

Test Code (run using ipython -pylab):

import scipy.sparse.linalg.eigen.arpack as arpack import scipy.sparse Ta=loadtxt("../../Matrix.dat") lam,ev=eig(Ta.transpose()) lam2,ev2=arpack.eigen(scipy.sparse.lil_matrix(Ta.transpose()),k=5,which="SM") print(sort(abs(lam))[0:5]) print(lam2)

"[ 0.00017072 0.00161566 0.00161566 0.00270811 0.00270811]" "[ 0.+0.j 0.+0.j 0.+0.j 0.+0.j 0.+0.j]"

[kyleabeauchamp]

Test Case Matrix