Kivancsiva tettel, igy en is csinaltam egy gyors osszahasonlitast, de egy szabadon elerheto projekttel (protobuf-2.6.0), hatha valaki meg akarja ismetelni. Minden tesztnel ugyanaz a szcenario futott le: tar, configure, time make -j. Az eredmenyek alapjan valoban megfigyelheto egy kis lassulas, de persze ez nem egy korrekt teszt, pusztan megismeteltem egy masik projekttel es kicsivel tobb mintaval azt, amit te csinaltal ;)
clang-3.4
real 63.73 user 114.66 sys 4.46
real 64.14 user 115.08 sys 4.30
real 64.03 user 115.03 sys 4.38
real 64.29 user 115.29 sys 4.51
real 64.47 user 115.50 sys 4.58
real 64.42 user 115.25 sys 4.65
real 64.56 user 115.58 sys 4.43
real 63.86 user 115.08 sys 4.43
clang-3.5
real 65.46 user 117.78 sys 4.57
real 65.91 user 118.08 sys 4.38
real 65.57 user 117.86 sys 4.58
real 65.59 user 117.57 sys 4.59
real 65.26 user 116.91 sys 4.75
real 65.57 user 117.39 sys 4.70
real 65.07 user 116.64 sys 4.59
real 64.99 user 116.62 sys 4.70
gcc-4.8.2
real 85.79 user 155.58 sys 7.18
real 86.48 user 156.04 sys 7.58
real 85.94 user 155.13 sys 7.54
real 86.01 user 155.51 sys 7.38
real 85.55 user 154.34 sys 7.77
real 85.49 user 154.74 sys 7.10
real 85.10 user 154.18 sys 7.26
real 85.46 user 154.73 sys 7.06
Es peldakent, hogy nem mindig a "make -j" a nyero megoldas ;)
clang-3.4: make -j3 (nproc+1)
real 59.72 user 109.39 sys 4.00
real 59.32 user 108.97 sys 4.04
...
clang-3.5: make -j3
real 60.51 user 110.94 sys 4.34
real 60.64 user 111.01 sys 4.29
...
gcc-4.8.2: make -j3
real 80.79 user 145.42 sys 7.62
real 80.31 user 145.66 sys 6.52
...