[comp.sys.apollo] Spice2g6 for apollo

bhv@areaplg2.corp.mot.com (Bronis Vidugiris) (04/16/91)

I have a version of spice2g6 source for the apollo with the following rather
cryptic comments:

I'm not quite sure where it came from, and if it's the 'best' version to
use.  I'm especially concerned about conversion #17 - combinging the
commons.  Is this the 'Adus' spice? Is this version/ conversion significantly
better for performance?  The conversion  (dated as 8/4/83) was done well
before sr10.

....

           Spice2g.6 8/4/83
This spice is supplied on an unsupported basis only.


This spice does not support 'distortion' analysis, due to time required to convert
the 'disto' routine.  It is included in 'src' if anyone wants to try...

.....

To convert Berkeley SPICE2G.6 from Unix on a VAX to Apollo we did the following:

1) Split routines
2) Change a compile error in errmem due to 'O' format specifier
3) Change a compile warning in dctran due to non-character format specs avfrm and avhdr
4) Same as (3) for dcop
5) Same as (3) for noise
6) Same as (3) for ovtpvt
7) Comment out IPOSTP line in spice.ftn
8) Delete the '_' on the end of the unix.c file
9) Add DFLOAT.FTN routine
10) Remove call to CLSRAW routine in dctran
11) Remove call to CLSRAW routine in acan
12) Remove '#define VAXUNIXASM' from unix.c
****
for general shipment (sans unix requirement), does not understand ipostp, whatever that is
13) replace xdate,xtime with mdate, mclock routines
14) replace unix zero, move, copy, locf routines
15) delete dblsgl and fwrite references in acan and dctran
16) replace second.ftn with an apollo version
17) combine common blocks into one for the loader to get variables on 4byte boundaries (largely
    by accident).  (DONT GET CAUGHT by /DC/ common exceeding 72 columns...)
    This involves fixing all duplicated names in common and in local variables.
    all are caught by the compiler except 'iord' in dcsol, 'ltab' in readin, 
    and 'vdsat' in dcop.  (I think...)
-- 
* Disclaimer *

This posting (probably) represents what the NNTP socket was told, but it
isn't representative of Company Policy or Opinion.

bhv@mot.com