Re: Garbage collection for system programming

Richard Xian <richard.xian@gmail.com>
1 Mar 2005 15:49:49 -0500

          From comp.compilers

Related articles
Garbage collection for system programming sheath1@gmail.com (Simon) (2005-02-28)
Re: Garbage collection for system programming torbenm@app-4.diku.dk (2005-02-28)
Re: Garbage collection for system programming Martin.Ward@durham.ac.uk (Martin Ward) (2005-02-28)
Re: Garbage collection for system programming richard.xian@gmail.com (Richard Xian) (2005-03-01)
Re: Garbage collection for system programming nmm1@cus.cam.ac.uk (2005-03-01)
Re: Garbage collection for system programming sheath1@gmail.com (Simon) (2005-03-01)
Re: Garbage collection for system programming torbenm@app-4.diku.dk (2005-03-04)
Re: Garbage collection for system programming pault@dessci.com (Paul Topping) (2005-03-04)
Re: Garbage collection for system programming nmm1@cus.cam.ac.uk (2005-03-04)
| List of all articles for this month |
From: Richard Xian <richard.xian@gmail.com>
Newsgroups: comp.compilers
Date: 1 Mar 2005 15:49:49 -0500
Organization: Compilers Central
References: 05-02-104 05-02-110
Keywords: GC, practice
Posted-Date: 01 Mar 2005 15:49:48 EST

Yes, if one individual processor is dedicated to an incremental GC,
the user program can be kept out of the interference from the GC. But
to a uniprocessor system, making GC incremental is not enough to
guarantee the time constraint of hard real-time tasks of the user
program, and rigorous schedulability analysis is required. Various
concurrent GC proposals can be the references.


On 28 Feb 2005 19:49:03 -0500, Martin Ward <Martin.Ward@durham.ac.uk> wrote:
>
> On-the-fly garbage collection algorithms have been around since at
> least 1978 where the garbage collector runs as a parallel process to
> the main program. ("On-the-fly garbage collection: an exercise in
> cooperation", E.W.Dijkstra et al, CACM Vol 21, Issue 11, Nov 1978, pp
> 966-975).
>
> With a multi-ocre CPU it should be possible to do garbage collection
> with practically zero overhead.



Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.