Floating-point numbers

Programs for article Floating point tricks:

• float2int.c --- conversion float to integers
• round2.c --- rounding double precision numbers

Other programs:

• round.c --- rounding numbers
• range01.c --- checks whether number lies in range [0, 1] (integer operations are used)

---

Type make to build float2int and round2; type make test to run performance tests.

Below are results from my Core i5 M540.

float2int:

function FPU, iterations = 10000, size = 65536                  0:01.43
function simple (C), iterations = 10000, size = 65536           0:01.26
function simple (x86), iterations = 10000, size = 65536         0:00.85

round2:

stdlib, iterations = 10000, size = 65536                        0:15.12
FPU FRNDINT, iterations = 10000, size = 65536                   0:18.43
FPU FISTP/FILD, iterations = 10000, size = 65536                0:01.01
simple method (C impl.), iterations = 10000, size = 65536       0:01.01