Hi All,

What header or compile-time trick allows one to determine the byte-
order of target CPU?

TIA,

-Le Chaud Lapin-

Le Chaud Lapin wrote:

AFAIK, you should use predefined macros like the ones here:

http://predef.sourceforge.net/prearch.html

There are runtime tricks but that is pretty silly unless you are using
an autoconf like thing in your build system to determine endianness.

On May 9, 8:25 pm, Le Chaud Lapin gmail.com> wrote:

Not possible. Depending on the CPU and the memory system, the byte
order may be dynamic, or it may vary among sections of memory. Or it
may be "mixed endian" (not strictly big or little), or have no
endianness.

clemd@acm.org wrote:

Wow. Do you have an example?

Sohail Somani wrote:

You will find that some hardware is switchable, or that the endianness
of integers and floats are not the same, or that the endianness is not
at the byte level. :-)

http://en.wikipedia.org/wiki/Endianness

Bo Persson

clemd@acm.org wrote:

Sohail Somani taggedtype.net> wrote:

The Digital PDP-11 and VAX systems (= arguably the most popular and
influential minicomputers in the world for the 1960s-70s and 70s-80s
respectively) were little-endian for integers, but stored floating-point
numbers in memory in a mixed-endian order. A floating-point number
(either IEEE or one of the VAX-only formats) which might be the register
bit-pattern 0x87654321 would be stored in memory with bytes in the order
(if my slight dim memories haven't faded too badly) first 2, then 1, 4,
3, 6, 5, 8, and finally 7. The reasons for this go back to backwards
compatability with early PDP-11 systems with had 16-bit memory buses.

On May 10, 1:08 pm, Sohail Somani taggedtype.net> wrote:

Dynamic: HP PA/RISC can switch endianness dynamically.

Vary among sections of memory: in the Intel i960CA, the endianness of
16 memory sections were individually selectable. I don't remember
whether they were dynamic.

clemd@acm.org wrote:

The MIPS chip that was in the SGI's and DEC pre-Alphas among others
could be switched. We referred to the DEC version as the SPIM.

clemd@acm.org wrote:

Correct me if I'm wrong, but doesn't the standard just say sizeof(int)

guarantee that a long long is any particular size...

Andy

In article
u12g2000prd.googlegroups.com>,
<"clemd@acm.org"> wrote:

Might be off topic, but where does c99 state range of long long
larger than range of int? My reading of the c99 standard states the
same thing as c89 and hence c++98 [and probably later revisions]
regarding sizeof various signed integral types. That is sizeof(X) ==
1 for any integral type X, is legal.
So that if the machine is a 'natural' 32 bit word machine, then
CHAR_BIT == 32 is legal.