general clarification#5
All floating point calculations should be done with double precision. This is particularly important to the litmus problem.