23 Mar
2007
23 Mar
'07
10:54 a.m.
On Thu, 22 Mar 2007 19:20:51 -0700, "Chuck Guzis" <cclist at sydex.com>
wrote:
On 22 Mar 2007 at 14:59, Steve Wilson wrote:
> >4900796 (Somebody will understand this reference)
>
> Someone is trying to boot Fortran off of the Disk on an IBM 1620 me
thinks...
Isn't 00796 where all-well-behaved programs went to die? I.e. reload
Monitor (sort of a CALL EXIT)?
--Chuck
We have a winner.
Just to add to the trivia, to boot the disk you needed a few more characters:
3400032007013600032007024902402111963611300102
Which if you were on a model 2, you needed to make sure that indirect
addressing was turned on (nothing worked right if you didn't!).
After doing that a few times, you remembered it pretty well.
OB Fortran reference:
There were several "load & go" Fortrans available for the 1620. The one I
liked the best was Witran. It fit into a 20k machine, but was interpreted.
Still not too bad for a Fortran. The best was the Monitor II Fortran II. It
generated automatic floating point instructions, and used index registers (only
for the compiler I think). Pretty speedy. The hardware floating point on a
1620-II was faster than the software on the (next generation) IBM 1130 (I used
that too). The 1620's bonus was that standard precision (in floating point)
was 8 significant digits. If you really felt ambitious, you could go MUCH
higher (28 digits with Fortran-II). Few machines today can match that in hardware!
--
Tom Watson
tsw at johana.com
____________________________________________________________________________________
Get your own web address.
Have a HUGE year through Yahoo! Small Business.
http://smallbusiness.yahoo.com/domains/?p=BESTDEAL
23 Mar
23 Mar
12:22 p.m.
On 23 Mar 2007 at 10:54, Tom Watson wrote:
We have a winner.
Just to add to the trivia, to boot the disk you needed a few more characters:
3400032007013600032007024902402111963611300102
Which if you were on a model 2, you needed to make sure that indirect
addressing was turned on (nothing worked right if you didn't!).
Never got to use a Model II; only used Cadets, (but with indirect
addressing and automatic divide). Ran Monitor I, which loaded pretty
much the same way--although my boot card had a "5" instead of a "1"
after the branch to 2402--read JCL from cards right away instead of
the typewriter.
I wrote an awful lot of SPS; probably more than FORTRAN. But FORTRAN
IID was far more of a "real" FORTRAN than the card versions--you
could have functions and subroutines and even call overlays. But the
floating point subroutines for SPS programs could accommodate 45
digits of mantissa--that'd be what--about 135 bits of binary
mantissa? Slow as molasses in January, but still pretty cool.
(I see you found my "Bottles of Beer" program!)
Cheers,
Chuck
After doing that a few times, you remembered it pretty well.
OB Fortran reference:
There were several "load & go" Fortrans available for the 1620. The one I
liked the best was Witran. It fit into a 20k machine, but was interpreted.
Still not too bad for a Fortran. The best was the Monitor II Fortran II. It
generated automatic floating point instructions, and used index registers (only
for the compiler I think). Pretty speedy. The hardware floating point on a
1620-II was faster than the software on the (next generation) IBM 1130 (I used
that too). The 1620's bonus was that standard precision (in floating point)
was 8 significant digits. If you really felt ambitious, you could go MUCH
higher (28 digits with Fortran-II). Few machines today can match that in hardware!
--
Tom Watson
tsw at johana.com
____________________________________________________________________________________
Get your own web address.
Have a HUGE year through Yahoo! Small Business.
http://smallbusiness.yahoo.com/domains/?p=BESTDEAL
6484
days inactive
6484
days old
1 comments
2 participants
participants (2)
-
cclist＠sydex.com -
sdc695＠yahoo.com