On 10/6/06, Xiao-Feng Li <[EMAIL PROTECTED]> wrote:


>A major design concept in this revision that was not made very clear
>in previous submissions is "space".  "GC" is a concept of a full (or
>standalone) garbage collector design. One GC can have multiple spaces,
>each of which can employ different "collection policy". With this
>design idea in mind, the previous trace-forward GC is adapted to be
>the collection policy for the nursery space, and the mark-compaction
>GC is adapted to be the collection policy of the mature space.

>As an embodiment of the design idea, GCv5 eliminates the global gc
>variable in existing Harmony GC, and different space managements only
>need to know a general gc data structure, instead of the generational
>gc structure.


 Thanks for the code Xiao Feng. This ( above ) lose coupling is a great
idea. We have discussed a little about space/collection policies/inter-space
relations on a previous thread.

This revision has been tested on Windows with SPECJVM98. As the first
>actual generational GC, any comments (or questions) are welcome!


 I will try to run the DeCapo suite against this.

 Thanks again,
 Rana

Reply via email to