2.2.3 Running under Win 95
- To: mathgroup at smc.vnet.net
- Subject: [mg2902] 2.2.3 Running under Win 95
- From: "Harry A. Smith" <hasmith at garnet.acns.fsu.edu>
- Date: Tue, 9 Jan 1996 02:17:08 -0500
- Organization: CMDS News machine
I had purchased the "standard" version of Mma for windows and was
informed by Wolfram that the standard version had been discontinued and
that for $145 I could upgrade to the enhanced or professional version.
So I did. I recently got the "enhanced" version expecting it to run
faster. Much to my surprise, it runs *slower* than the standard version!
Here is the timing test I did:
Plot3D[Sin[x]*Cos[y],{x,-Pi,Pi},{y,-Pi,Pi}]
On my Pentium 120, the standard version takes 1.92 seconds to execute
this. The "enhanced" mode takes from 2.8 to 3.6 seconds! There seems to
be no consistency on the timing redults I get. I am running both codes
under Win95 on a P120 with 16 MB RAM.
Interstingly, when I run the same code on a Mac Quadra 650 (68040, 33
mHz) using the enhanced version I get the same timings as on the Pentium
120! What's going on here? Did I waste my $145? My problem is that I use
this machine to perform computationally intense simulations and I just
can't afford this kind of performance hit when I "upgrade".
Other problems I have noted:
a. I can't interrupt any calculation! 2.21 works, 2.23 doesn't!
b. I can't load a file from the numbered file list
c. The timing function doesn't work anymore after a loop calculation, so
I can't time my code to test the timing of the calculations (which I
need to report in my research)
I did install correctly the patch to let Mma 2.23 work under Win 95. My
real question is why would the co-processor version actually run slower
than the non-coprocessor version? Did I do something wrong? Has anyone
else noticed this? Is there anything I can do?
--
Harry A. Smith
Professor of Biological Sciences
Tallahassee Community College
hasmith at garnet.acns.fsu.edu
"If you're not simulating, you're not stimulating!"
==== [MESSAGE SEPARATOR] ====