Path: utzoo!utgpu!news-server.csri.toronto.edu!bonnie.concordia.ca!uunet!indetech!vsi1!zorch!mykes From: mykes@zorch.SF-Bay.ORG (Mike Schwartz) Newsgroups: comp.sys.amiga.misc Subject: Re: How do we change the scheduler? (Was Re: Multitasking at home...) Message-ID: <1991Jan18.062608.14969@zorch.SF-Bay.ORG> Date: 18 Jan 91 06:26:08 GMT References: <7511@sugar.hackercorp.com> <1991Jan14.200405.19816@ncsuvx.ncsu.edu> Organization: SF-Bay Public-Access Unix Lines: 58 In article <1991Jan14.200405.19816@ncsuvx.ncsu.edu> cocoiii@hobbes.ncsu.edu (John Vestal) writes: > > I've been watching this very interesting discussion about >multitasking, and would like to make a few comments. I don't know >about how the kernals in MAC OS, AmigaDOS, OS/2, UNIX, or OS-9 work, but >I've used them all but OS/2. The Mac does a decent job of 'kludging >multitasking'. You can't (or I haven't found a way) of changing tasks >priority's on the Mac though. AmigaDOS does multitasking very >smoothly, but I have seen (mainly games) that will not let anything >else run. Well, at least not get too them, I don't know that they >aren't running, but to me they're not there. If you change a very CPU >intensive task to a higher priority, other tasks can 'stop'. On UNIX >you can run a lot of programs and you can never 'stop' a process, >without killing it, or setting it's priority as low as possible. The >difference between the two, if I have a compile that (for fun) takes >all of the CPU, and a shell running. The compile is one above the >shell. On AmigaDOS, the shell gets no time, until the compile is >done. On UNIX, the shell gets a certain percentage less then the >compile. It still gets some though. This is also how OS-9 works. >Tasks age, and after they are so old, according to the priority, they >get a CPU cycle. They also get any 'leftover' CPU, but AmigaDOS does >this too. Now, for me as a user, I want any process that I am running >to get some time. I should tell it at least how often I want it to >run. It can run more often, but at least this often. OS-9 does this, >and it doen't give up 'real-time response'. As a matter for fact, it >the OS of choice for realtime work. At least from what I've seen and >heard. Anyway that's my 2 cents worth. > > cocoiii (John Vestal) >-- >****************************************************************************** >* cocoiii@catt.ncsu.edu Phone: (919) 831-2890 North Carolina State * >* John Vestal P.O. Box 21537 Raleigh, NC 27607 University * >****************************************************************************** Games on the Amiga shouldn't multitask if they need speed (i.e. high speed animation, etc.) because the normal overhead of a multitasking system is enough to steal a lot of your CPU time. Also, games for the Amiga need to run on a 512K Amiga 500 (the most popular Amiga) and there typically is not even room for the OS and a decent sized game (or the game would suffer). Also, games are pirated heavily, and in a multitasking environment it is easy to run a debugger and the game at the same time and patch out copy protection and other forms of protection. Also, many of the images and sounds in a game are copyrighted or licensed material and hackers have no business gaining access to them. As far as the Amiga tasking system and high priority tasks stealing all the CPU time... When you write a program for the machine, you can decide what priority your program and its associated tasks run at. It is considered POOR programming practice to hog the CPU in any way shape or form. However, if you really want to do some CPU intensive processing at a high priority, you do have the power to do it. And there are some cases when it is desireable, like when refreshing your editor's screen. This can make the editor seem much more responsive, even though it can slow down that compilation you have going in the background. But the machine responds immediately to user input, which can be very user friendly. mykes