py26-numpy is much slower than NumPy compiled for MacPython

Konrad Hinsen konrad.hinsen at fastmail.net
Wed Dec 8 08:02:26 PST 2010


On 08.12.2010, at 08:21, Konrad Hinsen wrote:

> Update: I profiled both runs using Shark, and found that when using the MacPorts version, 57% of the total time is spent in function feclearexcept. In the MacPython run, this function doesn't even show up. This makes me suspect that the difference is in the handling of floating point numbers, specifically floating-point exceptions.

Second update: feclearexcept is a function in numpy/fenv/fenv.h declared as __inline. Apparently it has been inlined in the official NumPy binaries, but not in the MacPorts installation.

Konrad.


More information about the macports-users mailing list