🚀 go-pugleaf

In <1992Sep11.123326.22143@tc.cornell.edu> bai@msiadmin.cit.cornell.edu (Dov Bai-MSI Visitor) writes:

>In article <92Sep10.221229.26167@acs.ucalgary.ca> bauwens@acs.ucalgary.ca (Luc Bauwens) writes:

>>
>>Now about Paranoia:
>>
>>For MS Fortran 5.1, I get *exactly* the same behavior for an
>>OS/2 16 bits executable run under OS/2 2.0, for a DOS executable
>>under OS/2 2.00 and for the DOS executable under DOS 5.00.  I
>>don't have a machine with OS/2 1.3 handy, so I didn't try that.

>Few things about the MS-Fortran experiments:

>1) I found that the OBJ file produced for the DPARA.FOR is larger than
>   64K (~70k). This may cause failure of the executable. I split
>   DPARA.FOR into 3 files. There are many subroutines so it is easy
>   to split.

>2) I found that compiling with optimization (the default)  causes 
>    failures with DPARA. When  I added the -Od switch the problems
>    went away.


>>Luc B
>>

>Dov


The spara program fails with Watcom 8.5 version of the Fortran compiler,
running in a DOS box under OS/2. The double precision version passes 
with 1 defect. Has anyone run these tests on Watcom 9.0 Fortran? Does
this Single-precision, Double precision dichotomy still hold? 

I hope they have fixed up the floating point code. Its useless to have
a Fortran compiler thats deficient in the one area that the Language
is the undisputed leader in (at least in terms of historical inertia
and available code). I wonder how many people program Windows using
Fortran, so in my opinion all that effort would probably be better
spent in cleaning up their floating point libraries. Their 
educational price is great though, so I can't gripe much. Had I been
a shop that had purchased their compiler to do industrial 
strength numerical work, I'd be hopping mad. 

I'd run out and buy the Lahey compiler, except that its 4 times as
expensive and doesn't support OS/2 2.0 yet. 

Appreciate any comments on all this.


Gyan P. Sinha
UBC, Vancouver.