Hello, It seems that for the cpp back-end, the activity-local object is allocated in a thread-local free-list. And the list will be mark and sweep. The allocate and GC need lock, so have some synchronous overhead.
However, how does the Java back-end implement activity-local object allocation? Does Java back-end also has the GC synchronous overhead which may affect scalability? Thank you very much. Regards, -Tetsu ------------------------------------------------------------------------------ This SF.net email is sponsored by Make an app they can't live without Enter the BlackBerry Developer Challenge http://p.sf.net/sfu/RIM-dev2dev _______________________________________________ X10-users mailing list X10-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/x10-users
