I doubt its the GC kicking in, but you could run it with the following environment variable set just to ensure that its not doing anything.

    GOGC=-1

EDIT: Trying it out quickly shows a small improvement actually, but so small as to likely be noise as I was doing other things on the machine.

    Summary
      ./mainnogc 1000 ran
        1.01 ± 0.06 times faster than ./maingc 1000

You've set the Go code to use int64 but the other faster implementations are using int32.