<HTML><BODY>Hi, Neks!<br><br>I am using a my_full_restart subroutine as in cyl_restart example. In logfile it seems that it works well, because Nek reads first 3 steps from previous calculation, but I also write to the file the meaning value of friction velocity (u_tau) and it differs. During first 10-20 dimensionless time there is a huge peak in the instant profile of u_tau. Then it begins to oscyllate near an exact meaning, but such behaviour looks strange in case of full_restart using. Do you have any ideas why is it so?<br><br>Best regards, <br>Vladislav</BODY></HTML>