Path: utzoo!mnetor!tmsoft!torsqnt!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!rpi!crdgw1!crdos1!davidsen From: davidsen@crdos1.crd.ge.COM (Wm E Davidsen Jr) Newsgroups: comp.arch Subject: Re: OS/2 is dead? Message-ID: <3051@crdos1.crd.ge.COM> Date: 18 Dec 90 13:58:14 GMT References: <28775@usc> <14887@ogicse.ogi.edu> <3078@canisius.UUCP> Reply-To: davidsen@crdos1.crd.ge.com (bill davidsen) Organization: GE Corp R&D Center, Schenectady NY Lines: 32 In article pete@csc-sun.mckinsey.com (Peter Gaston) writes: | From personal experience: | - OS/2 appears slow, even on a 386 machine, but that may be more | an artifact of PM and the file system That's my impression. Running programs which have multiple processes and little screen activity seems to work reasonably well. Someone here measured total CPU spent in user processes as 85% of clock time under load, so the overhead is not much higher than numbers I see on a loaded UNIX system, also measured as "anything not given to the user is overhead." | - it supports both 'normal' and real-time priorities, is the | real-time as useful as it claims? It looks good as a process control o/s. Someone posted about comparing it with Venix in an application which needed real time data aquesition and control. I believe the evaluation showed both would do the job but the Venix user interface was better. If you don't need realtime capability, you may actually get an increase in kernel overhead caused by forcing preemptable points in the logic. I plan on trying V.4 capabilities in this area sometime before spring. | - writing a device driver is a bear I'm happy to say I don't know about that! -- bill davidsen (davidsen@crdos1.crd.GE.COM -or- uunet!crdgw1!crdos1!davidsen) VMS is a text-only adventure game. If you win you can use unix.