Re: [Jikesrvm-core] copyMS

[Steve B. <Ste...@an...>]

> I think it's OK to recommend it as the default; however, the lack of 
> compaction might be a cause for concern for any long-running apps (e.g., 
> a long-running eclipse session).   Comments?

Fair question.  The short answer is that I don't know.  However, this
MS design is intended to be low fragmentation by virtue of its
coalescing capacity---something I've put a fair amount of energy into.
Also by virtue of the filtering that happens in the copyMS "nursery",
fragmentation is much less of an issue than in MS.  As a side note,
fragmentation & compaction is something we're investigating at ANU (and
I know UMass is also looking at it, though from a different
perspective).

Cheers,

--Steve