g++ version is 5.3.0.
#include <algorithm>
int main() {
return 0;
}
test$ time g++ test.cpp
real 0m0.203s
user 0m0.073s
sys 0m0.031s
test$ time g++ test.cpp --std=c++11
real 0m0.761s
user 0m0.554s
sys 0m0.130s
Similar results over multipfle attempts. Only a minor slowdown (0.03s) with including the vector header.
The straight forward answer is that there is obviously more to compile. Using -Q -ftime-report GCC will print out profile statistics. Keep in mind that the section on the left is not reliable because it changes everytime you run the command, but the TOTAL is always consistent. For C++03 and C++11 respectively:
Execution times (seconds)
phase setup : 0.00 ( 0%) usr 0.00 ( 0%) sys 0.00 ( 0%) wall 1189 kB (16%) ggc
phase parsing : 0.03 (100%) usr 0.02 (100%) sys 0.06 (100%) wall 6301 kB (83%) ggc
preprocessing : 0.01 (33%) usr 0.01 (50%) sys 0.04 (67%) wall 488 kB ( 6%) ggc
parser (global) : 0.00 ( 0%) usr 0.01 (50%) sys 0.00 ( 0%) wall 3626 kB (48%) ggc
parser struct body : 0.02 (67%) usr 0.00 ( 0%) sys 0.01 (17%) wall 778 kB (10%) ggc
parser function body : 0.00 ( 0%) usr 0.00 ( 0%) sys 0.01 (17%) wall 436 kB ( 6%) ggc
TOTAL : 0.03 0.02 0.06 7558 kB
Execution times (seconds)
phase setup : 0.00 ( 0%) usr 0.00 ( 0%) sys 0.00 ( 0%) wall 1384 kB (11%) ggc
phase parsing : 0.05 (100%) usr 0.03 (100%) sys 0.10 (100%) wall 10953 kB (88%) ggc
|name lookup : 0.00 ( 0%) usr 0.01 (33%) sys 0.01 (10%) wall 1301 kB (10%) ggc
preprocessing : 0.01 (20%) usr 0.00 ( 0%) sys 0.03 (30%) wall 878 kB ( 7%) ggc
parser (global) : 0.01 (20%) usr 0.01 (33%) sys 0.01 (10%) wall 4592 kB (37%) ggc
parser struct body : 0.02 (40%) usr 0.00 ( 0%) sys 0.01 (10%) wall 2837 kB (23%) ggc
parser function body : 0.01 (20%) usr 0.01 (33%) sys 0.01 (10%) wall 478 kB ( 4%) ggc
parser inl. meth. body : 0.00 ( 0%) usr 0.01 (33%) sys 0.03 (30%) wall 937 kB ( 8%) ggc
symout : 0.00 ( 0%) usr 0.00 ( 0%) sys 0.01 (10%) wall 0 kB ( 0%) ggc
TOTAL : 0.05 0.03 0.10 12490 kB
That's a difference of 12490 - 7558 = 4942 KB, not that big of a difference. Now as to why there's such a big difference in GCC 5.3.0, it's a QoI issue that's probably been fixed in later releases.