Version française
Home     About     Download     Resources     Contact us    
Browse thread
practical functional programming
[ Home ] [ Index: by date | by threads ]
[ Search: ]

[ Message by date: previous | next ] [ Message in thread: previous | next ] [ Thread: previous | next ]
Date: -- (:)
From: Ken Wakita <wakita@i...>
Subject: Re: practical functional programming

In message (<3A0EF904.C65AC976@ozemail.com.au>)
from John Max Skaller <skaller@ozemail.com.au>,
talking about "Re: practical functional programming",
on Mon, 13 Nov 2000 07:09:40 +1100

skaller> Ken Wakita wrote:
skaller> > 
skaller> > A circumstance where reference counting outperforms modern trace-based
skaller> > collectors is where memory access cost is much higher than the
skaller> > conventional memory system and thus memory access required for tracing
skaller> > is much higher than the cost for counter maintenance.  One such example
skaller> > is distributed environment.  Another maybe systems with very, very slow
skaller> > memory such as file systems, persistent object systems, and PDAs.  I am
skaller> > curious if there are other circumstances using conventional memory
skaller> > system where reference counting is faster.
skaller> 
skaller> I don't know how a trace-based collector works. Can you explain?
skaller> [Does this have something to do with a write barrier on pointer
skaller> stores?]

Trace-based collector (or tracing collector) is a name given to a
class of garbage collection algorithms that trace the object graph in
the heap.  It covers most of the garbage collection algorithms,
mark&sweep and copying.  If you incorporate generational techniques
probably you need to use write barrier.

Ken Wakita
Tokyo Institute of Technology