[pypy-issue] [issue1279] Numerical code for list of floats 3x slower than for list of ints.

[Armin Rigo]

Armin Rigo <armin.rigo at gmail.com> added the comment:

It's an issue of a list of ints-or-floats: if you change the example to use
floats everywhere --- in the creation of A, in the creation of B, and in the
creation of result --- then the result is again fast, slightly faster than with
ints.

I would recommend that we look at implementing the notion of list of
32-bit-int-or-float, by using invalid float representation (NaNs) to mean "it's
really an int".

----------
nosy: +arigo

________________________________________
PyPy bug tracker <tracker at bugs.pypy.org>
<https://bugs.pypy.org/issue1279>
________________________________________