single precision:
        sizeof(REAL) =              4
        FLT_MANT_DIG =             24
        machine epsilon =           1.19209E-0007 [IEEE 754 32-bit macheps]           
        smallest positive number =  1.40130E-0045 [IEEE 754 smallest 32-bit subnormal]

double precision:
        sizeof(DOUBLE PRECISION) =  8
        DBL_MANT_DIG =             53
        machine epsilon =           2.22045E-0016 [IEEE 754 64-bit macheps]           
        smallest positive number =  4.94066E-0324 [IEEE 754 smallest 64-bit subnormal]

quadruple precision:
        sizeof(REAL*16) =          16
        LDBL_MANT_DIG =           113
        machine epsilon =           1.92593E-0034 [IEEE 754 128-bit macheps]                
        smallest positive number =  6.47518E-4966 [IEEE 754 smallest 128-bit subnormal]