MathGroup Archive 1992

[Date Index] [Thread Index] [Author Index]

Search the Archive

RE>Conflict between Mma and

  • To: mathgroup <mathgroup at yoda.physics.unc.edu>
  • Subject: RE>Conflict between Mma and
  • From: Roberto Sierra <73557.2101 at compuserve.com>
  • Date: 21 Aug 92 13:23:29 EDT

       Reply to:   RE>Conflict between Mma and Compact Virtual

[Regarding Larry Riddle's problems with Mathematica 2.0/2.1
when run using Compact Virtual (virtual memory) on a MacSE
and System 6.0.7]:

I can't say, with any certainty, what the problem *is*, but
can certainly propose some workarounds which may make your
problem disappear -- or at least change into a different
beast.

First off, I've been using Mathematica *a lot* with virtual
memory under System 7 on a IIx and haven't encountered any
vm-related problems (yet).  Other users say they have had
problems, but the concensus seems to be that the crashes
are occasional, and not always repeatable (which is why
Wolfram has not addressed some of those issues.)  Please,
Mathgroup, correct me if I'm wrong about that.

It's possible that the problem lies with Compact Virtual,
possibly having to do with a weird accelerator interaction.
You might find the problem goes away if you use System 7's
virtual memory (then you *won't* want to install the
Compact Virtual INIT).

[Larry -- I will dig back and forward you the message
thread from Mathgroup that dealt with vm-related issues
under System 7 so you'll know more about what I'm saying
But give me a couple of days -- I'm swamped at work right
now and don't have the time to go back through the archives.]

This solution may or may not work on an SE with 68030
accelerator, depending on the nature of the Apple ROMs.
I *know* that this would work on an SE/30, IIx, IIcx, etc.,
because I've done it before, and it *may* also work on
a II with an MPPU installed, though it depends on the
ROMs, once again.

1)  Start by calling the Mobius accelerator support people
    and ask them if anyone's used that product under Sys 7
    on a Mac SE.  Then ask if MODE32 works under System 7
    -- in other words, if it's possible to use 32-bit
    addressing modes on their accelerated SE.  [You need
    32-bit addressing to run virtual memory under Sys7]
    
    If so, you should be set to go, and things may just
    work (or at least fail in new and interesting ways.]
    
    If not, there's probably not much use in continuing,
    since you'll need MODE32 to run Virtual Memory
    under System 7 in the first place.


2)  Back up your current 6.0.7 (just in case) and install
    Sys7 on top of that.  OR, set up Sys7 on a separate
    hard disk and boot from that disk instead.


3)  Install Apple's MODE32 control panel.  This can be
    ftp'd from ftp.apple.com, and is available on a zillion
    bulletin boards, CompuServe, etc.  The same is true for
    System 7.0.1 (the current version) if you don't already
    have a copy.  If you can't find these things anywhere,
    drop me a line and we'll make arrangements to e-mail or
    snail-mail the appropriate items.

4)  Open MODE32 control panel under System 7 and turn MODE32
    patch ON [takes effect after reboot].

5)  May have to reboot at this point -- [not sure].

6)  Open Memory control panel, enable 32-bit addressing and
    virtual memory, setting up a swap file of the appropriate
    size on the hard disk somewhere.

7)  Reboot.

8)  Try running Mathematica and see what happens.  If the
    system crashes now, it's most likely because of one of
    two things:

        SE ROMs are not compatible with MODE32.
        [SE/30, IIx, or IIcx would therefore be needed.]
        Mobius accelerator board is at fault.

    Either way you're in big trouble.  Time to buy a IIsi,
    Quadra, or fx if you can afford 'em.  Even a used IIx
    or IIcx or SE/30 would probably be totally sufficient.

    In general, I've had nothing but trouble working on
    machines with accelerators.  There's always some bizarre
    incompatibility lurking somewhere which will selectively
    trash certain applications (e.g., some MIDI stuff I was
    using.)  Not worth the hassle.  Sell your old machine
    buy a used SE/30 or IIx/cx and you'll have a much happier
    life.


Anyway, if you're able to get System 7 running, in 32-bit
addressing mode with Virtual Memory up, there's a good chance
you'll be all set.  Or, it's also possible that this is doomed
to failure.  Either way, you'll know more than you did when
you started.  Drop a line and let me know how it goes...


\\|//
 - -
 o o
  J   roberto sierra
  O   tempered microdesigns
 \_/  73557.2101 at compuserve.com









  • Prev by Date: Re: FontForm
  • Next by Date: Mma Solve[] oddity (a much shorter version)
  • Previous by thread: translating the APL 'deal' function into M recursively, procedurally and functionally - one-liners are faster
  • Next by thread: Mma Solve[] oddity (a much shorter version)