Path: utzoo!utgpu!news-server.csri.toronto.edu!bonnie.concordia.ca!uunet!zaphod!ncar!gatech!ncsuvx!news From: cocoiii@hobbes.ncsu.edu (John Vestal) Newsgroups: comp.sys.amiga.misc Subject: Re: How do we change the scheduler? (Was Re: Multitasking at home...) Message-ID: <1991Jan15.160832.18718@ncsuvx.ncsu.edu> Date: 15 Jan 91 16:08:32 GMT References: <17210@cbmvax.commodore. <7504@sugar.hackercorp.com> <42459@ut-emx.uucp> <7511@sugar.hackercorp.com> <1991Jan14.200405.19816@n Sender: news@ncsuvx.ncsu.edu (USENET News System) Organization: NCSU Computing Center Lines: 25 burley@geech.ai.mit.edu (Craig Burley) writes: >"Aging" low-priority processes is a useful feature, especially in >multi-user systems, but it can be critical for a real-time system to >allow specification of priorities such that a low-priority task NEVER >executes as long as any higher-priority tasks do. >I think, therefore, when you say "it doesn't give up 'real-time response'", >you mean "real-time feel", either that or the real-time demands you're >placing on the system are loose enough that an occasional scheduling of >a non-real-time low-priority task doesn't hurt. No, I didn't make my self clear here. You can establish a minumim process priority that is not aged, and a maximum process priority that gets the CPU before other processes. However, only process that NEED this kind of response get it, and I can change whether it gets that time or not. In AmigaDOS, you don't tell the computer... it tells you. I prefer the flexiblity of having the "both of both worlds", ie. realtime if you need it, aging overwise. -- ****************************************************************************** * cocoiii@catt.ncsu.edu Phone: (919) 831-2890 North Carolina State * * John Vestal P.O. Box 21537 Raleigh, NC 27607 University * ******************************************************************************