Greetings,
Nothing big here.
I was updating my view of multicore GC strategies and happened on a
paper with an interesting conclusion:
"Our evaluation suggests that today, there is no conceptual reason to
believe that the
pause time of a stop-the-world GC will increase with the increasing
number of cores and memory size of multicore hardware."
https://hal.inria.fr/file/index/docid/868012/filename/gidra13asplos-naps.pdfThe description in the paper sounds a lot less complicated than a number
of multicore gc strategies I have been looking at.
Also of note,
http://concurrencykit.org with a library of cross-platform
lock-free code, which might make playing with the above strategy a bit
simpler.
I really don't have time to chase this, but just in case someone out
there has a grad student looking for a project.. ;^)
FYI,
-KenD
-KenD