NEC-LIST: NEC4 with MS Fortran Powerstation

The Microsoft PowerStation version 1.00 compiler has some pretty serious bugs.
One is that if a double-precision complex variable is assigned the value of a
single-complex one or vice-versa (the change in precision being implicit), the
imaginary part isn't assigned as it should be. This leads to significant errors
with codes like the NEC-81 version of NEC-2, which has a mixture of single and
double precision complex numbers. NEC-4.1S contains only single-precision
complex numbers (and I presume that 4.1D has only double-precision ones), so
this problem doesn't show up with NEC-4.1. There are others, however, that do.

Microsoft produced a modified version, 1.0a, which fixed the bugs. It may not be
evident from the PowerStation sign-on screen that the revised version is
present; mine still says 1.00 even though I have the update. (There's also the
problem that the replacement disks Microsoft sent were version 1.00. It took a
second call and wait to get 1.0a.)

Here's how to tell which version you have. The following files are in the
F32\BIN directory:

  File v.1.00 v.1.0a

BINDMSF.EXE 73,206 01-30-93 64,302 01-15-94
F23232.EXE 404,480 02-08-93 489,984 12-22-93
LINK32.EXE 277,504 02-03-93 52,224 02-02-94
LINK.EXE ABSENT 385,536 01-19-94

A number of other files also changed. This is just a sampling for
identification. There may have been later updates than 1.0a.

I've found the following file to give an error when run on NEC-4.1 compiled with
PowerStation v. 1.00 but not on those compiled with 1.0a or other compilers. The
error, a divide-by-zero, occurs when writing or preparing to write the SOMS.NEC
or SOMD.NEC file. Be sure and erase any SOMS.NEC or SOMD.NEC file in the
directory before running the test, to force the program to write a new one. I
haven't attempted to track down the cause of the error.

The reaction of NEC-2 to this file probably depends on the version of NEC-2.
This file will only test NEC-4.1.

CM K5RP
CE
GW 1,7,0.,.1524,1.524,0.,0.,5.0292,.00103
GW 2,10,0.,.1524,1.524,0.,17.8917,1.524,.00103
GW 3,10,0.,0.,5.0292,0.,17.7393,5.0292,.00103
GW 4,6,-.4572,0.,1.524,-.4572,.1524,5.0292,.00103
GW 5,10,-.4572,0.,1.524,-.4572,17.7393,1.524,.00103
GW 6,10,-.4572,.1524,5.0292,-.4572,17.8917,5.0292,.00103
GW 7,6,0.,17.8917,1.524,-.4572,17.8917,5.0292,.00103
GW 8,6,0.,17.7393,5.0292,-.4572,17.7393,1.524,.00103
GE 1
FR 0,1,0,0,7.
GN 2,0,0,0,13.,.005
EX 0,1,4,0,1.414214,0.
RP 0,181,1,1000,90.,0.,-1.,0.,0.
WG NGFS.NEC
EN

I've had limited experience trying this test. Any results from others would be
appreciated.

Roy Lewallen
Received on Thu Dec 19 1996 - 21:48:00 EST