parallel search speedup/overhead

[flok]

new results:

Code: Select all

7 1 8 330537 21.529 15353.105114032236 D7-D5 => 25
7 2 8 2119386 67.214 31531.912994316663 D7-D5 => 25
7 3 8 287309 7.856 36571.91955193483 D7-D5 => 25
7 4 8 350920 7.231 48529.940533812754 D7-D5 => 25
7 5 8 305976 5.295 57785.83569405099 D7-D5 => 25

So in this case with 2 cores, well, strange things happen. This is not always reproducable. This also happens with other configurations.

For now I'm happy that the result move is always the same.