Re: [Chicken-users] static or dynamic nursery

[felix winkelmann]

On 6/3/06, Brandon J. Van Every <address@hidden> wrote:
> In Makefile.am it appears that nsample.exe benchmarks the Chicken build
> and hardwires some default nursery values deemed best for the platform
> it was built on.  This strikes me as unproductive when distributing a
> Chicken binary.  I'm building on a modest 866 MHz P-III, for instance.
> Granted we're not currently in the habit of distributing Chicken
> binaries, people compile from source, but that could change.  Or can
> it?  Is there some inherent reason that Chicken binaries can't be
> distributed, that source compilation is required?

One can skip the nursery sampling and just chose 64 or 128k
(which seems to be the optimum for current machines). It's not crucial,
but may give slightly better performance.

Building chicken from sources is still the recommended method - to use
chicken effectively, one has to work on that level anyway, IMHO.
That doesn't mean binary releases aren't useful, of course. Especially
bootstrapping issues can be handled better that way.


(felix)