help about gmf_t

[王渊]

I have a problem about gmf_t.
I need have 30 decimal point,but when it use the float division of gmp,the result is very different with the result of Windows' calculator and Java BigDecimal.
example:
3.14/5.789
the result of gmp:0.5424080152012437875651563769429
the result of Windows' calculator and Java BigDecimal:0.5424080152012437381240283295906
 
how can i get the correct result?
 
lei peng