[MLton-devel] IBM Research PL day

[Stephen Weeks]

> > >         Perry Cheng, IBM Research:
> > >         A Defragmenting, Mostly Non-moving Garbage Collector
> 
> Here are the notes I took during Perry's talk yesterday.

Thanks for the great notes.

> - Results
>   + worst case fragmenation never realized on Java SPEC benchmarks
>   + memory footprint: 1.1 to 1.4X live data
>   + read barriers practical with optimizations
...
>  prospects for MLton: the smaller memory footprint seems really attractive
>                       if we ever get around to an GTK or OpenGL interface,
>                         long GC pauses might be noticable
>                       redirection costs might be high, both space and time

The footprint is nice, but I'm not sure it's worth the cost.  OTOH,
the incrementality is essential in some situations.  The redirection
time cost is what really worries me, even with making redirection
explicit and doing CSE.  The upshot -- it looks like this is a ways
out there for MLton right now.

_______________________________________________________________

Have big pipes? SourceForge.net is looking for download mirrors. We supply
the hardware. You get the recognition. Email Us: bandwidth@sourceforge.net
_______________________________________________
MLton-devel mailing list
MLton-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/mlton-devel