Embedded and real-time JVMs (Was Re: [gclist] write-barrier implementation in Java VM?)

Manoj Plakal plakal@cs.wisc.edu
Mon, 6 Nov 2000 13:57:16 -0600

Sylvain Louboutin wrote (Mon, Nov 06, 2000 at 11:51:08AM -0800) :
> just out of curiosity,  what is the size of the heap for KVM (typically)?
> i.e.,  is it worth bothering with an incremental GC for a really tiny
> VM? (which I understand is the whole idea behind KVM, correct me if
> I am wrong)?
> I found out (the hard way...) that a plain and simple stop-the-world
> approach can be surprisingly effective with small heaps...

	Not to start the whole ref-counting thread all over again :),
	but are any embedded JVMs considering the use of reference
	counting as an incremental technique?

	Lucent's Inferno OS has a VM called Dis, which uses
	ref counting as the primary technique with some form
	of tracing collection as backup. The Limbo language 
	compiles to the Dis VM and has constructs for creating
	acyclic structures which don't leak when ref-counting.
	But cyclic structures are allowed too.

	Just wondering if other embedded systems use it too.

	Also, I hear about proposals for GC in real-time JVMs,
	but isn't it a bit of an oxymoron to have "real-time"	
	and "GC" next to each other, at least for tracing GCs
	whose run-times are difficult to bound? I guess it
	does matter on the definition of "real-time".