<HTML><BODY>Hi, Neks!<br><br>Could I set tolerances less than 10^-8 in solvers via .par or .rea file, or 10^-8 is the minimum? In documentation "double precision" means 8 digits, but could Nek5000 write files with 16 digits?<br><br>Best regards,<br>Mark<br></BODY></HTML>