From ... From: Erik Naggum Subject: Re: the evil of continuations Date: 2000/06/21 Message-ID: <3170590392650958@naggum.no>#1/1 X-Deja-AN: 637259592 References: <3949E0DC.8E428904@eurocontrol.fr> <3dmdy739.fsf@alum.mit.edu> <4Pu25.84$4p1.1925@burlma1-snr2> <394A7C71.F1F75FE7@acm.org> <394A860B.360DFBE@acm.org> <871z1wqqzz.fsf@orion.dent.isdn.cs.tu-berlin.de> mail-copies-to: never Content-Type: text/plain; charset=us-ascii X-Complaints-To: newsmaster@eunet.no X-Trace: oslo-nntp.eunet.no 961602205 8834 195.0.192.66 (21 Jun 2000 15:43:25 GMT) Organization: Naggum Software; vox: +47 8800 8879; fax: +47 8800 8601; http://www.naggum.no User-Agent: Gnus/5.0803 (Gnus v5.8.3) Emacs/20.6 Mime-Version: 1.0 NNTP-Posting-Date: 21 Jun 2000 15:43:25 GMT Newsgroups: comp.lang.scheme,comp.lang.lisp * Michael Sperber | Unfortunately, OS threads are often *way* too heavyweight for the | really interesting applications of threads. * Tim Bradshaw | unfortunately, unless you use the OS-supplied threads you'll never run | on a multiprocessor. Whatever happened to good old interprocess communication, shared memory, etc, on multiprocessor systems? Whatever it is people are doing with threads can be handled without involving any kernel. There need be no particularly complicated implementation, either. At issue, however, is whether it makes sense to write languages and runtime systems that maintain the threads and surrounding support or hack things by hand in what passes for the "popular" languages. I'm sorry, but I completely fail to understand why OS threads is the only way to utilize "modern" multiprocessor systems. Maybe I'm just too old to be able to pretend we didn't have SMPs in the 1980's. #:Erik, sighing -- If this is not what you expected, please alter your expectations.