This is the mail archive of the cygwin@sourceware.cygnus.com mailing list for the Cygwin project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: Re: long long vs long



In a message dated 7/24/98 1:24:02 AM, smorris@xionics.com wrote:

>Octal makes sense in an environment with
>word sizes of multiples of 3 bits. Hex is only useful with multiples of 4
>bits.


I found hex useful on 36-bit word machines like GE600/Honeywell6000 where
floating point sub-fields were aligned on 4-bit boundaries, but of course
characters were aligned either on 6-bit or 9-bit boundaries.  It wasn't
difficult to have hex display software recognize patterns which made more
sense in character or octal.  So why did octal persist so long on 16/32 bit
machines?

>The IBM 360 was 36 bits.

The  704 and 7094 were 36 bits, as were some would-be competitors of the 360.
These competitors apparently believed, mistakenly, that IBM customers would
switch brands rather than switch word lengths, and that the superiority of
36-bit binary floating point over 360-style 32-bit hex would carry the day.
As it turned out, of course, prices of 32-bit memory dropped so fast that
64-bit double precision was more affordable than 36-bit single with extended
registers.
-
For help on using this list (especially unsubscribing), send a message to
"gnu-win32-request@cygnus.com" with one line of text: "help".


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]