home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.os.os2.misc
- Path: sparky!uunet!sybus.sybus.com!myrddin!tct!det
- From: det@tct.com (Dennis Taylor)
- Subject: Re: Mix of 3-chip and 9-chip simms with OS/2
- Message-ID: <2B35D039.663F@tct.com>
- Date: Mon, 21 Dec 1992 14:10:01 GMT
- References: <8163@lib.tmc.edu> <BzE1Kz.A8F@news.cso.uiuc.edu> <8168@lib.tmc.edu>
- Organization: TC Telemanagement, Clearwater, FL
- Lines: 27
-
- In article <8168@lib.tmc.edu> jmaynard@oac.hsc.uth.tmc.edu (Jay Maynard) writes:
- >In article <BzE1Kz.A8F@news.cso.uiuc.edu> nap42487@uxa.cso.uiuc.edu (Nishith A. Patel) writes:
- >>Well, if you first say that the RAM is fine even for heavy duty diagnostics,
- >>but not for OS/2, then say the problem is with RAM, there is ome kind of
- >>contradiction. It is not like most RAM out there is "mortal" RAM, and there
- >>is some very privileged, select few RAM chips that are the ultimate in RAM
- >>chips. All RAM that is in perfect condition are pretty much the same, and getting
- >>RAM that happens to "work" with OS/2 is pure luck.
- >
- >No, that's not what I'm saying. I'm saying that it's been known to happen that
- >the RAM had defects that only OS/2 had uncovered; the diagnostic programs are
- >not as good a diagnostic as simply running OS/2. I don't know why OS/2 is such
- >a stress test, but it's generally agreed that OS/2 is a better diagnostic for
- >memory than diagnostic programs.
- This is not new with OS/2. I experienced the same problems with
- mini-computers in the early '70s. We would experience mysterious crashes
- (unfortunately we did not have parity checkers), run diagnostics and turn
- up nothing. The vendor did have a "multitasking" diagnostic that ran several
- tests concurrently, including disk I/O, which was more reliable than other
- diagnostics, but not as reliable as the OS with real applications.
-
- The problem is not finding memory that pleases OS/2, but memory that meets
- the manufacturer's hardware specifications, even under the stressful
- situations implied by multitasking CPU, disk, and display operations.
-
- Dennis
-
-