------- Comment #3 from rguenth at gcc dot gnu dot org 2008-02-28 16:47 -------
With the patch proposed for PR34043 we get
f:
.LFB2:
addss (%rdi), %xmm0
movd %xmm0, (%rsi)
ret
instead of
f:
.LFB2:
addss (%rdi), %xmm0
movss %xmm0, -4(%rsp)
movl -4(%rsp), %eax
movl %eax, (%rsi)
ret
so expansion can handle
f (a, b, e)
{
<bb 2>:
*b = VIEW_CONVERT_EXPR<int>(*a + e);
return;
}
well compared to
f (a, b, e)
{
union a c;
<bb 2>:
c.f = *a + e;
*b = c.i;
return;
}
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33989
