[Fwd: Re: Parrot GC?: Will donate real-time GC code.]
Kurt Stephens
-------- Original Message --------
Subject: Re: Parrot GC?: Will donate real-time GC code.
Date: Thu, 17 Apr 2003 16:20:40 -0700 (PDT)
From: Ask Bjoern Hansen <a...@perl.org>
To: Kurt Stephens <k...@kurtstephens.com>
On Thu, 17 Apr 2003, Kurt Stephens wrote:
Hi Kurt,
It sounds very interesting!
You should post your message on perl6-i...@perl.org.
- ask
> I have written parts of a conservative, non-copying "treadmill" GC. It
> threads GC work while the mutator requests allocations using a write
> barrier and four-color nodes segmented into pages chained by size and/or
> type. It has special support for large and medium objects to reduce
> fragmentation. I had started implementing a write barrier through
> OS-level memory protection faults, but never finished it. It will
> require work, but it has, IMHO, some potental.
>
> It can collect dead objects without stopping the world or requiring a
> separate collector thread. It can be tuned to guarantee that collection
> phases during allocations will complete in a specific amount of time
> (assuming that the system has memory to spare until the next round).
>
> I was wondering how far along your GC work for Parrot has gone and if
> the Parrot architecture will be open for other GCs, like the Boehm GC or
> UMASS Toolkit.
>
> If you are interested in it, it is located at
> http://kurtstephens.com/research/index.html .
>
> I am willing to donate it to Parrot (as it is collecting dust).
>
> I've been using perl for 12 years and I think that reference counting is
> annoying and completely unnecessary considering the amount of excellent
> MM work that has been done over the last 15 years.
>
> I haven't had much time to look at Parrot, but since I'm looking for a
> job, I have plenty of time now. :)
>
> Thanks,
> Kurt Stephens
> ks....@kurtstephens.com
> http://kurtstephens.com
>
> References:
>
> "A Memory-Efficient Real-Time Non-Copying Garbage Collector", Tian F.
> Lim, et. al., ACM SIGPLAN vol. 34, Num. 3, p. 188, March 1999. ISMM '98
> Procs. of the ACM SIGPLAN International Symposium on Memory Management,
> Oct 17-19, 1998.
>
>
>
>