[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: continuation efficiency

From: Martin Grabmueller
Subject: Re: continuation efficiency
Date: Sun, 08 Jul 2001 09:59:49 +0200

> From: address@hidden (Thomas Bushnell, BSG)
> Date: 07 Jul 2001 16:11:29 -0700
> Martin Grabmueller <address@hidden> writes:
> > call/cc does a full copy of the machine stack, to be able to install
> > it again when the continuation is invoked.  This requires memory
> > allocation and copying, which is a bit slow.
> > 
> > call/ec, on the other hand, does not need to preserve the stack, since
> > it only allows calls upwards in the call chain.  Think of call/ec as
> > throw/catch and you'll see.  So call/ec is less powerful, but faster
> > (in Guile, other Schemes may vary).
> This all depends on the *implementation*.  The original post implied
> that call/ec was not a primitive.  If it is implemented as a wrapper
> around call/cc, then how can it be more efficient?

Yes, but aren't we talking about the implementation in Guile here?
IIRC, you started this thread, asking about the performance of call/cc
in Guile, and Marius posted an example of call/ec, based on
catch/throw, _not_ based on call/cc.  It was this implementation I was
talking about, which uses catch/throw---these are primitives, and
they are faster than call/cc *in Guile*.

If I mixed something up here, please clear it up.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]