Re: [Gretl-devel] RNG (was: Can't open datafile)
-------- Original-Nachricht --------
Datum: Wed, 19 Jan 2011 07:30:11 +0100 (CET)
Von: "Riccardo (Jack) Lucchetti" <r.lucchetti(a)univpm.it>
Just crossed my mind: what happens to random number generation on machines
which don't support sse2 (which I assume is the case on non-x86
architectures)?
I guess you're not including x86-64 in "non-x86", so in general you're
probably right.
we fall back on the old Mersenne Twister. But if this is the case,
we'd
have that scripts with the same seed would produce architecture-specific
results. Do we want this?
Yes I think we do want this. I would argue that nowadays the vast majority of gretl users
are running on sse2-capable hardware, so this affects only very few settings. (Things may
change if people want to run gretl on smartphones and such, but maybe even those have
sse2? And do people really want to do that?) Of course, I'm all for including
informative and visible warnings in the gretl output when such a fallback occurs.
cheers,
sven