Bytes order and words order

Gabriel Dos Reis gdr@integrable-solutions.net
Fri Aug 23 06:22:00 GMT 2002


Nathan Sidwell <nathan@acm.org> writes:

| Gabriel Dos Reis wrote:
| 
| > | Don't assume that the word-order of a floating-point number will be the
| > | same as the word-order of an integral type.  It isn't the case on the FPA
| > | co-processor for the ARM for instance.
| > 
| > OK, thanks.  I was actually thinking of the word-order of
| > floating-point numbers not that of integral types.  I think I should
| > name it  __TARGET_FLOAT_WORDS_ORDER__ ?
| Why TARGET?

If the library is run on a machine that is not the target machine, the
there are some support codes that depends on the host (the running machine)
and the target (the simulated machine).

| Surely __FLOAT_WORD_ORDER__ is sufficient. I'm a bit confused
| about what exactly you mean by word order for a floating point type.

Assume a 32-bit machine on which a double and a long long are 64 bits
wide -- thus implemented as two words aggregates.  The issue 
is what is the "order of the words" in a double (floating-point
word-order) and in a long long (integral-type word order). 

| But you and Richard seem to know what you're talking about.

I'm sorry I gave the impression of a private chat.

-- Gaby



More information about the Gcc mailing list