> In message <19991105033224.A2612@uni-koblenz.de>, Ralf Baechle writes:
> +-----
> | > Both macros do the same thing on all *currently supported*
> | > architectures. One never know when a funny thing from Mars shows up;
> | > perhaps a 48-bit architecture, who knows...
> |
> | Honeywell had 9-bit based machines, that is longs / pointer were 36 bit.
> +--->8
>
> PDP-10 had 36-bit words as well. But I don't think they subdivided into
> 9-bit chunks (although they did subdivide into 6-bit chunks for i.e.
> filenames, which were stored in SIXBIT).
They subdivided into bytes that were anywhere from 1 to 36 bits wide,
depending on what you needed. TOPS-10 used SIXBIT for filenames; TOPS-20
used ASCIZ strings (five bytes/word with the LSB unused). Some tape
formats were buffered in 9-bit bytes. PDP-11 based FEPs often exchanged
buffers of 8-bit bytes. I think tht 36-bit word width stems from a
decision to use existing 18-bit memory boards in the PDP-6. The original
design used 18-bit addresses.
IIRC people are still building PDP-10 clones, so this is not quite as OT
as you might think -- somebody might want to port Linux to 'em someday.
(Usually the layering goes the other way: there's a PDP-10 emulator that
runs on Linux. :-)
-- Mark H. Wood, Lead System Programmer mwood@IUPUI.Edu Please, no more software products offering a "richer experience"! I have indigestion of the brain already. Give me a more ascetic experience.
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/