[Numpy-discussion] Segmentation fault when dealing with larga rank arrays

Francesc Alted falted at openlc.org
Wed Nov 6 08:16:08 EST 2002


Hi,

I'm having some segmentation faults when dealing with large rank arrays in
Numeric.

You can see some examples in the following:

>>> zeros((1,) * 21) == zeros((1,) * 21)
Segmentation fault

Of course, this kind of usage is quite unusual, but I'm interested on it
because I'm developing tests for a package based on Numerical and wanted to
check the larger rank supported.

>>> ones((1,) * 41)
Segmentation fault

In this case, I know that Numeric supports array ranks until 40, but in
opinion, an error should be raised instead of merely givin a "Segmentation
fault".


I'm having these problems using Numeric 20.2.1, 21.0 and 22.0 in both Python
2.1.x and 2.2.x. My platform is Intel under Linux Debian 3.0.

Thank you,

-- 
Francesc Alted                            PGP KeyID:      0x61C8C11F
Scientific aplications developer
Public PGP key available: http://www.openlc.org/falted_at_openlc.asc
Key fingerprint = 1518 38FE 3A3D 8BE8 24A0  3E5B 1328 32CC 61C8 C11F




More information about the NumPy-Discussion mailing list