On Fri, Dec 20, 2013 at 5:30 PM, Dave Angel <da...@davea.name> wrote: > On Thu, 19 Dec 2013 16:32:37 +0100, Wolfgang Keller <felip...@gmx.net> > wrote: >> >> With Windows it *is* "normal". An experienced software developer >> once even explained the reason to me. When a single process on > > Windows >> >> does I/O, then the system essentially falls back to "single > > tasking". >> >> Or (non-)"cooperative multitasking" at best, depending on how >> dissocial the developer of that process is. > > > If you were told this 20 years ago, perhaps. But Windows hasn't been > running on DOS for a long time. Starting with NT 3.1, that's nonsense.
Actually, one thing I might believe is that when one process trips a page fault and has to be swapped back in, all other processes freeze, waiting on the memory manager. But I'd want to see some evidence that that's the case. It's plausible (based on some of the problems I've seen) but by no means necessarily true. ChrisA -- https://mail.python.org/mailman/listinfo/python-list
