At 12:55 AM 2/17/2004, you wrote:
At the dev con, I was reminded that in ARM code, not only had 2-byte
integers be aligned on 16-bit boundaries, but 4-byte integers also had
to be aligned on 32-bit boundaries. What about 8-byte integers? On
64-bit boundaries? Does this apply to floating point data as well, eg.
4-byte floats on 32-bit boundaries and 8-byte doubles on 64-bit
boundaries?

The ARM4T architecture only requires 4-byte alignment. On some ARM cores with FP support, double math is faster with quantities aligned on an 8-byte boundary, but there shouldn't be any access issues.


Lastly, does ARM little endianness affect floats and doubles or are
these endianness-independent formats?

Yes, these format are endian-reversed from 68K.


(BTW, it was good to meet you last week! Your app really is impressive.)



--
For information on using the Palm Developer Forums, or to unsubscribe, please see 
http://www.palmos.com/dev/support/forums/

Reply via email to