MathGroup Archive 1995

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

Search the Archive

Re: MMA for OS2/Windows front end, Help

  • Subject: [mg2190] Re: MMA for OS2/Windows front end, Help
  • From: roussel at physics.mcgill.ca (Harold Roussel)
  • Date: Fri, 13 Oct 1995 06:24:30 GMT
  • Approved: usenet@wri.com
  • Distribution: local
  • Newsgroups: wri.mathgroup
  • Organization: McGill University Computing Centre
  • Sender: daemon at wri.com ( )

edg at wam.umd.edu (Ezra Greenberg) wrote:

>I upgraded a while back to MMA for OS2 and have not been able to get the 
>windows front end to work in a satisfactory manner.  On my machine, the TCP 
>link is oh so very slow.  The PM front end, is, well, useless as far as I can 
>tell.  I have many complicated notebooks, and I have yet to figure out a way 
>to work with the various parts in a reasonable way.

>Are there any other options?  Is anybody else out there having problems with 
>Mathlink to the windows front end?

For some reason, and I'll probably complain to Wolfram about that, the
windows front-end keeps polling the keyboard like mad.  The result of
this is that, when the mathlink communication is established, the
front-end sends data to the os/2 kernel which only gets a fraction of
the cpu time to do its work.  That's the reason why it is so slow.

Now, for the solution (or at least interim fix), you must first get a
utility that boost the priority of your os/2 programs like "nice" (try
hobbes.nmsu.edu in dir /os2/unix, filename nice10.zip).  Once you have
it, go in your wnmath22 directory.  There should be a file called
math.cmd.  In this file you will find the line (with a different drive
possibly),

start /PGM "h:\math\mathpm.exe " %1 %2 %3 %4 %5 %6 %7 %8 %9

replace it with

start h:\apps\bin\nice.exe /c h:\math\mathpm.exe %1 %2 %3 %4 %5 %6 %7
%8 %9

Replacing the paths as appropriate.  Beware though that the
mathematica kernel will run at time-critical priority!  This means
that if your calculation is going to take a long time, your system
might be unusable.  The best thing to do would be to start MMA in
normal priority, with delta 32 (the front-end runs in normal priority,
delta 1).  That would probably give it enough CPU time over the
front-end to complete calculations in a reasonable amount of time, yet
the desktop could remain responsive.  I seem to remember that there is
another "nice" program that can adjust the deltas.  You should
probably look on hobbes.

In my case I never need the windows front-end so I'm just using a
plain text like interface.  If I need a scroll-back buffer I can
always use emacs, or use ckermit and telnet into my own machine in
loopback mode.  The best solution would be a native front-end, or at
least a better windows version.  I have to send an email to Wolfram...

>Ezra D. Greenberg
>Department of Economics
>University of Maryland
>edg at wam.umd.edu





//
// Harold Roussel            email: roussel at physics.mcgill.ca
// Ph.D. student             phone: (514) 398-6506
// High Energy Physics
// McGill University, Montrial, Quibec
//
// Using OS/2 Warp...and still waiting for NT 3.51.
//
// "...sous cette peau de mital je dipose devant toi tout ce qui me
// reste; muscles, os, sang, et le coeur qui le met en mouvement."
//                    Lancelot au Roi Arthur, dans Excalibur.




  • Prev by Date: Re: Inverse Z-Transform and/or Long Division
  • Next by Date: Re: Why gives Plot[GTK,{q,o,30}] negat. min. at q=15 ? (GTK:=1/3 q^2 - 10 q + 100)
  • Previous by thread: Re: MMA for OS2/Windows front end, Help
  • Next by thread: Re: MMA for OS2/Windows front end, Help