[Openmcl-devel] Quick HW question...

[David L. Rager]

+1 for the idea that caching across multiple cores affects CCL and
SBCL performance.  I have some tests that demonstrate this, if
anyone's interested.

> CCL uses spinlocks on most platforms; acquiring a spinlock involves a
> loop like:
>
> (block got-lock
>  (loop
>    (dotimes (i *spin-lock-tries*)
>      (if (%store-conditional something somewhere)
>        (return-from got-lock t)))
>    (give-up-timeslice)))

Mind if I ask which platforms use spinlocks?  I'm interested in
knowing about linuxx86, linuxx86-64, powerpc, powerpc64, and
darwinxx86-64.

Thanks!