News for hackers. Stuff that matters.
Re:The Future of the LispM by the_why_of_y on Sunday February 08, 2015 @11:16AM (#9017576)
GC as such requiring read barriers in the mutator is nonsense.

You may need to require read barriers if you do concurrent compaction of the heap, as a lot of the "real-time" collectors do.

But state of the art collectors like G1 can do even concurrent marking (but stop-the-world compaction) without mutator read barriers.

http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.63....

There are no comments on this comment yet.