http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51838
Bug #: 51838
Summary: Inefficient add of 128 bit quantity represented as 64
bit tuple to 128 bit integer.
Classification: Unclassified
Product: gcc
Version: 4.7.0
Status: UNCONFIRMED
Severity: enhancement
Priority: P3
Component: c
AssignedTo: [email protected]
ReportedBy: [email protected]
void foo(__uint128_t *x, unsigned long long y, unsigned long long z)
{
*x += y + ((__uint128_t) z << 64);
}
Compiles into:
mov %rdx,%r8
mov %rsi,%rax
xor %edx,%edx
add (%rdi),%rax
mov %rdi,%rcx
adc 0x8(%rdi),%rdx
xor %esi,%esi
add %rsi,%rax
adc %r8,%rdx
mov %rax,(%rcx)
mov %rdx,0x8(%rcx)
retq
The above can be optimized into:
add %rsi, (%rdi)
adc %rdx, 8(%rdi)
retq
