Andrew Hutchinson wrote:
Hi all,
just spend some days going over gcc-avr and missed optimizations.
Which is a huge bunch of work! Thanks!
<snip>
The result is a 1.25% saving in code size for a simple mixed
application. Pretty good for such a simple change!
Very good! What test cases were you using as a test case?
1.25% code size is very significant, since it translates into a speed
improvement as well, so 1.25% reduction of an inner loop can buy even
more of a speed up.
For more floating point, the saving might well be higher as it demands
more contiguous 32 bit registers.
That makes sense. I don't have any good test cases for 32 bit AVR code.
But actually, simply measuring libm is probably a very good test case in
a practical sense.
-dave
_______________________________________________
AVR-GCC-list mailing list
[email protected]
http://lists.nongnu.org/mailman/listinfo/avr-gcc-list
