On my system there isn't a noticeable difference in execution times. However the disassembly suggests (at least at first glance) addition of
aa
and
bb
should be faster than that of
a
and
b
:
pro.c += pro.a + pro.b
000000e0 mov eax,dword ptr [ebp-78h]
000000e3 mov eax,dword ptr [eax+0Ch]
000000e6 mov edx,dword ptr [ebp-78h]
000000e9 add eax,dword ptr [edx+4]
000000ec mov edx,dword ptr [ebp-78h]
000000ef add eax,dword ptr [edx+8]
000000f2 mov edx,dword ptr [ebp-78h]
000000f5 mov dword ptr [edx+0Ch],eax
cc += aa + bb
00000197 mov eax,dword ptr [ebp-14h]
0000019a add eax,dword ptr [ebp-0Ch]
0000019d add eax,dword ptr [ebp-10h]
000001a0 mov dword ptr [ebp-14h],eax