stan wrote:
On Fri, 31 Jul 2009 20:12:05 +0100 (BST)
Patrick Dupre <pd520(a)york.ac.uk> wrote:
> Hello,
>
> I do observe abnormal results with my applications
>
> Moving from a 32 to 64 bits architecture, using, c, C++, perl, is
> they something that I should be aware of ? like size of the
> float, integer ?
>
Yes, in C and C++, shouldn't matter in perl. If you search on 64 bit
compatibility you will find articles. There was a post on this list a
long time ago about it also.
Basically, there are special names that adjust depending on arch size.
Some of the type sizes are different between 32-bit and 64-bit
architectures, even between different 32-bit architectures. Other types
are fixed, no matter what architecture you use.
A char, for instance, is always 8 bits or one byte. A Unicode char, if
implemented, is 16 bits or two bytes.
An int is a "natural" integer, a.k.a. "word", for the CPU
architecture,
so 8, 16, 32 or 64 bits depending on the CPU.
More detail can be found here:
http://en.wikipedia.org/wiki/C_variable_types_and_declarations, among
other places.
Cheers,
--
Paul