What makes GHC's GC good for purely functional workloads?

[Antoine Latter]

On Wed, Nov 23, 2011 at 7:55 AM, Antoine Latter <aslatter at gmail.com> wrote:

>
> It sounds like the new "Garbage First" collector might be closer to
> these goals than the older HotSpot JVM collectors:
>
> http://labs.oracle.com/jtech/pubs/04-g1-paper-ismm.pdf
>
> It was built with different goals, but the end result is that it (can)
> focus on collecting garbage while its young.
>
> The paper was written a few years before the first version shipped, so
> I'm not sure how the implementation differs.
>

Although it looks like the young-generation collections pause for
quite a bit longer than 10us, and if I'm reading the paper correctly
don't necessarily collect the entire young generation.