Floating-point and mixed-endianness in D-Bus (was: dbus mini-summit)

[Rémi Denis-Courmont]

Le mardi 9 août 2011 22:20:41 Havoc Pennington, vous avez écrit :
> I don't think there needs to be a performance issue for the
> conditional conversion. It does not need to be more than one
> conditional per message in theory, which will not be even remotely a
> measurable overhead.

I don't really see how you'll make this into a single check, unless maybe you 
just convert the whole message buffer to native endianess at the time of 
reception.

Sure, you could check once and then select a table of function pointers based 
on that one-time check. But function pointers are just as slow as branching, 
probably even slower, as they break prediction. Furthermore, you'll get the 
function call overhead.

> If going to fixed byte order, then I'd probably go with little endian,
> because it seems sucky to have the swap overhead in the 95% case
> instead of the 5% case.

Sure. Little endian is not unique to CPUs :-)

Or there is also the option of not interoperating buses of different 
endianess. Even in the rare networked D-Bus case, that mixed endianess should 
be an exception.

-- 
Rémi Denis-Courmont
http://www.remlab.net/
http://fi.linkedin.com/in/remidenis