[Gc] What is the overhead of -DALL_INTERIOR_POINTERS?
hans.boehm at hp.com
Wed Nov 18 13:00:37 PST 2009
I think most experience has been that in the expected case it doesn't make much difference. But the probability of occasional misbehavior (i.e. significant retention of non-garbage) is significantly higher with -DALL_INTERIOR_POINTERS. Note that you also have GC_register_displacement(), which may be a better way to inform the collector about peculiarities of your language implementation. Or not. It really depends.
> -----Original Message-----
> From: gc-bounces at napali.hpl.hp.com
> [mailto:gc-bounces at napali.hpl.hp.com] On Behalf Of Lothar Scholz
> Sent: Wednesday, November 18, 2009 12:31 PM
> To: gc at napali.hpl.hp.com
> Subject: [Gc] What is the overhead of -DALL_INTERIOR_POINTERS?
> I'm designing a new compiler/language which will use the GC.
> What is the overhead of -DALL_INTERIOR_POINTERS instead to
> assure that pointers are always pointing to the origin?
> Is it worth to implement some language features differently
> if this avoids -DALL_INTERIOR_POINTERS?
> Best regards,
> Lothar Scholz mailto:scholz at scriptolutions.com
> Gc mailing list
> Gc at linux.hpl.hp.com
More information about the Gc