[SciPy-User] ANN: SciPy 0.10.1 release candidate 1

Pauli Virtanen pav@iki...
Tue Feb 14 14:21:55 CST 2012


14.02.2012 20:49, Samuel John kirjoitti:
[clip]
> Just the known arpack related scipy failuers and umath_complex related failures numpy failures.

The bad news is that we though the ARPACK issues were fixed :/

Could you paste the relevant pieces from the test report? The error is
probably with test tolerances when doing iterative single-precision
inverses inside iterative eigenvalue calculation, but I'd like to double
check.

Thanks,

-- 
Pauli Virtanen



More information about the SciPy-User mailing list