17 Oct
2007
17 Oct
'07
2:37 a.m.
On 10/17/07, woodelf <bfranchuk at jetnet.ab.ca> wrote:
Ethan Dicks wrote:
It's kinda worse than that. The FORTRAN-II compiler compiles down to
rather ordinary and expected sequences of PDP-8 machine instructions.
The FORTRAN-IV compiler, AFAIK, really just compiles down to sequences
of FPP-12 instructions, which makes it fast if you have an FPP-12, but
really slow (because of emulation) if you do not.
There's a bunch of register-level information on the FPP-12 in the
1977-1978 edition of the "Small Computer Handbook" (it has a PDP-8/a
on the cover, naturally). I received a copy of that from the local
DEC field service guys back in 1982 when I picked up my first PDP-8.
Unfortunately, I had an -8/L, and in my "n00bness", failed to
appreciate all that was in there. Much later, when I found a 1968
copy of the handbook, much more made sense. In the meantime, I did
read about the FPP-12 and decided that it would be cool to play with,
but I've never run across one.
-ethan
... FORTRAN-IV (slow on an SBC-6120
since it doesn't have the optional-for-a-PDP-8/a FPP12 math
co-processor)...
I still find it hard to belive that PDP-8/a had a math co-processor.
That does make the PDP-8 the smallest computer with real floating
point I can think of. Ben.