Return to BSD News archive
Xref: sserve comp.unix.pc-clone.32bit:4133 comp.unix.bsd:12470 comp.os.linux:53570 comp.unix.questions:38037 comp.os.mach:3201 comp.unix.solaris:5731 biz.sco.general:7352 Newsgroups: comp.unix.pc-clone.32bit,comp.unix.bsd,comp.os.linux,comp.unix.questions,comp.os.mach,comp.unix.solaris,biz.sco.general Path: sserve!newshost.anu.edu.au!munnari.oz.au!constellation!osuunx.ucc.okstate.edu!moe.ksu.ksu.edu!vixen.cso.uiuc.edu!howland.reston.ans.net!europa.eng.gtefsd.com!uunet!hobbes!timr From: timr@sco.COM (Tim Ruckle) Subject: Problems on Fast Machines (was: Unix close for 486 - commens requested) Organization: The Santa Cruz Operation, Inc. Distribution: world Date: Fri, 20 Aug 1993 19:23:23 GMT Message-ID: <1993Aug20.192323.16423@sco.com> References: <CBAs9D.MH4@murdoch.acc.Virginia.EDU> <hastyCBvJrI.CMy@netcom.com> <CByvHr.AMJ@egr.uri.edu> Sender: news@sco.com (News admin) Lines: 55 In article <CByvHr.AMJ@egr.uri.edu> black@cs.uri.edu (John Black) writes: } It's possible to have too much machine. Hi John, what I would say is that if you live out on the bleeding edge of hardware technology, it's probable that you will run into difficulties. At least this has always been my experience. I think it is because the commercial software must always play catchup to the accelerated capabilities of the hardware (since you can't really test on it until it's already here) and software engineers cannot or do not always anticipate conditions which will exist a couple years down the road, when the project they're working on now will actually be released and being used in the field... } I'm sitting on a Gateway 2000 4DX2-66V that was purchased to run LynxOS. } Its got a fancy VESA local bus, fast hard drive, zippy video card, and } runs Windows 3.1 like greased lightning. It can't even load Lynx though, } because...well, nobody really knows. If I cripple the machine by diabling } cache, turbo, IDE block mode, etc., it will sometimes boot Lynx, but } usually not. I'm not sure about LynxOS, but I can speak to what may be going on with the SCO systems. } Further, the June '93 issue of Byte magazine reported on "fast 486 machines" } and their ability to run SCO UNIX--several of them had problems similar to } mine, and in at least one case the solution was to cripple the machine } as I've had to do. On SCO UNIX System V/386 Release 3.2 Operating System Version 4.0, and SCO Open Desktop Release 2.0 systems there is a loop in the Adaptec 154x driver ("ad") and Adaptec 174x driver ("eiad") which can execute too quickly on a fast machine (many 486/66DX2 and some 486/50DX machines). The driver asks the adapter whether it is present, then checks 131,072 times for a response. The check executes faster on "fast" machines, and can time out before the host adapter responds. When that happens, the operating system assumes that there is no adapter present. SCO provided fixes (SLS unx365 for UNIX 3.2v4.0 and 3.2v4.1 and SLS oda366 for Open Desktop 2.0) for this, and fixed it in the current releases. You could also "cripple" the machine during installation, and then patch the kernel afterwards so that you could boot and run at full speed. I know of no other problems with "fast" machines on SCO systems other than this--do you recall which version(s) the Byte folks were using in their report? I don't think it's possible to *ever* have too much machine, you just need somebody who is able to catch up when the iron sprints out ahead. ;^) -timr -- I think there's a world market for about 5 computers. --Thomas J. Watson, Chairman of the Board, IBM (circa 1948)