Hello, What is the highest floating point precision supported by gfortran on a IA-32 architecture (ie P4)? Is it possible to get to KIND=16? Thanks, Jerry (Still learning every day)