Byte sizes (was Re: 2.8M 3.5' floppy

woodelf bfranchuk at
Mon Mar 14 11:27:51 CST 2005

der Mouse wrote:

>Well, it sounds reasonable, but probably isn't the most useful, as it
>would mean that, for example, the PDP-8 had 12-bit bytes.
Umm words...   it has six bit bytes.  I think the lack of 18 bit byte 
cpu's has limited
the use of 9 bit bytes. ( No comment  from any  36 bitters )

>>The PDP-10 is an excellent example of when this isn't true.
>>The smallest addressable unit is a word, which is 36 bits.
>>A byte is, as noted, anything between 0 and 36 bits.  Bytes are
>>stored in a word, as many as can be fitted.  To access bytes on a
>>PDP-10, you have a byte pointer, which consist of a word address, and
>>a bit pointer, and byte size.
>That sounds a whole lot like a hardware-supported way of addressing an
>object of an arbitrary size in bits.  And that would mean that bytes of
>any size *are* individually addressible.
>Or have I misunderstood?
I belive so, but I have not programmed a 10. I suspect this is in 
regards to string operations
rather say accessing character sized data data. Two different things.

>>The fact that people today seem to believe that byte addressable is
>>the only possible thing, along with a byte being 8 bits, is plain and
>>simply because they haven't seen any other.
Well in most cases a byte is a unsigned 1/2 half word. I still like view 
of the PDP-11 a byte
is signed data. I think the lack of real byte access has forced the C 
standard to have unsigned
bytes because the machines ( 8080,Z80) can't handle real bytes .

>>There even are relatively modern machines which don't fit it very well.
>>Some DSPs, for example, have 32 bits as their smallest directly
>>addressible unit.
It is word access I found the problem , not byte acess in the smaller 
machines a problem.
As for the 386(+) I can say it still needs 8080 style code generation 
for simple compilers
and compilers provide it. IE pesudo code.
Ben alias woodelf

More information about the cctech mailing list