PDP-11 firsts? - test-drb@ccmp.vtda.org

9 Nov 2010

On 11/06/10 02:03, ard at p850ug1.demon.co.uk  (Tony Duell) wrote:

...

 >>>> >  >>>  >    > 
 The Philips P800 series has the PC as a general register (register 0).
>>> >  >>  >
>>> >  >>  >    Could you use it like any other register?
> >  >  There were some restrictions. I don't think you could shift it (at
least
> >  >  ont on the P850). But for many instructions it was just another
register.
 >
 >  Cool. So you could add to it, index by it, and so on?  I think so. I
can't find my programemrs pocekt guide fo the machine at
 the momnet. But I am pretty sure that register 0 could often be used like
 any other register. 
Ok. Nice.

...
  There is at least one oddity. There are no
autoicrement/autodecrement
 adressing modes on the P800s. However, in some cases, if you use the PC
 in some instructions it is autoincrements. For example the 'immediate'
 addrtessign mode isessentialy a (PC) operation (i.e. take the word
 pointed to by register 0) and has that bit pattern. But for obvious
 reaosns the PC is autoincrements then. 
The PDP-11 actually pulls that trick on a few addressing modes as well.
Addressing mode 6 and 7 autoincrement the register if it is the PC, but 
not for other registers.

(So a thing like MOV FOO,R0 is encoded as
  MOV x(PC),R0
  .WORD .-FOO

and the PC should obviously be incremented again here, but x(Rn) does 
not exist as a variant with autoincrement of the register.)

...
  > >  >
 I guess the P800 wasn't totally ortogonal but it was a lot more
> >  >  orthogonal than many other machines.
 >
 >  Indeed sounds nice. When did the machine appear?  My P850 CPU service manual
is copyright 1972, which alas puts it after
 the PDP11, but I thoguht the series was around a bit before that. 
Hmm. Maybe still a first for the PDP-11 then...? :-)

...

>>>> >  >>>  >    >   What do you mean by condition codes
here?
>>> >  >>  >
>>> >  >>  >    The four low bits of PSW.
> >  >  Err... I don;t think that's helpful. Quite a lot of machines with a
> >  >  status register have a 'low 4 bits' of it:-). But that
doens't make them
> >  >  condition codes. Similarly uf you hapopen to implement the same
> >  >  functionality using other bits of a status registers, doesn't that
make
> >  >  them condition codes?
> >  >
> >  >  What I was asking was what fucntionality do you require of these
> >  >  condtiion codes other than there being conditional jumps on carry, zero,
etc?
 >
 >  Sorry. I was just being lazy, and trying to explain condition codes by
 >  referring to what they are on the PDP-11.
 >
 >  To try and be more specific then: condition codes are bits that are
 >  set/reset as a result of operations performed by the processed, and upon
 >  which you can the make conditional branches/jumps on.  Now the P800 has a
very odd way of doing this. There is a 2 bit status
 registers. It is set differnet ways according to the results of some
 instructiuos (for example, an arithmetic instruciton will set it one way
 if the result is 0, a differnet way if there's a carry out, etc). I/O
 operations set it one way for device ready, another for certain errors, etc.

 The condtional branches have a 3-bit condition field. You can branch on
 the status beits being any particualr value (00, 01, 10, 11), them not
 being one of 3 values 9I forget which one is omitted) or 'always' 
Sounds slightly different than the condition codes of the PDP-11 then, 
but not really like the immediate test and do instructions like the 
PDP-10. Closer to the PDP-11, it would seem.

	Johnny