Path: utzoo!utgpu!watserv1!watmath!att!occrsh!uokmax!apple!usc!cs.utexas.edu!uunet!smosjc!brett From: brett@smosjc.UUCP (Brett Coon) Newsgroups: comp.sys.amiga Subject: Multitasking is slower Keywords: games, multitasking Message-ID: <279@smosjc.UUCP> Date: 3 Jun 90 22:09:03 GMT Organization: S-MOS Systems, San Jose, Ca., USA Lines: 26 >From: Tim Friest - Programmer/Analyst >The point the people are making about games needing to multi-task is simply, >if I run the game without anything else running, it is exactly the same as >if the game took over the machine and didn't let me run anything else.... NO! For a program to allow multitasking, it has to be "nice". It must own and disown the system resources (esp. blitter) using OS calls, and do so often enough to allow other tasks a chance to use them. Depending on the program and the specific circumstances, this additional overhead could make a real difference in performance. And I'm not defending poor programming practices like skipping the OS altogether, either. The point is, if a program doesn't have to worry about giving any other programs use of the system resources, it can take control of them all, and not release control until it's finished. Whether it accesses them directly or through the OS is another issue. Personally, I agree that multitasking games are preferable to non-multitasking ones. And I don't feel that copy protection or programmer laziness are good reasons to abandon multitasking. But given the choice, I would always pick a good non-multitasking game over a poor multitasking one. So, while I think it's good to stress the importance of multitasking to the game programmers on the net, I also think there are some instances in which multitasking support isn't worth the price. Luckily, the price is small, so these instances SHOULD be rare. And games should DEFINITELY be HD installable.