“…Core 1 continues to complete the instructions in the pipeline. When the trace runs 10,10,10,11,12,13,13,14,16,16 History length (bits) 4,9,13,24,37,53,91,145,256,359 to the instruction that the migration finish (we call this instruction Migration Over Point -MOP) on core 1, core 2 waits core 1 finish the remaining instructions and returns registers, dirty cache blocks. Then core 1 stop running trace, core 2 continues to run the remaining trace.…”