This is the mail archive of the guile@sources.redhat.com mailing list for the Guile project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: OK, what about some resolution (Re: GUILE's GC - why we struggling to solve already solved problems?)


>>>>> "Michael" == Michael Livshin <mlivshin@bigfoot.com> writes:

    Michael> * current Guile GC sucks.

Really?  That's news to me.  The current evidence is that GGC is 50%
faster than BGC.  Obviously BGC can be tuned and made more precise, it
will always need to scan more than GGC.  

I guess it's worth pointing out that BGC uses a VM barrier to
implement generational GC.  VM barriers were considered for
Generational GC in guile and discarded because they were far too
inefficient.  The trade off is that accessors must be provided which
implement a software write barrier for any guile (ie smob) type.
Since accessors for internal guile objects are a Very Good Thing, this
requirement is also A Good Thing.  In anycase, it only affects the small
amount of glue code needed to add a new type to guile.  (BGC can also
use software marks, but that has the same problem as in Guile).

Cheers,

Clark

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]