Page 8 - PowerPoint Presentation

PENGKODEAN FLOATING POINT

Representasi floating point IEEE

 V = (–1) x M x 2 E
s

Kode biner floating point

s exp frac

1 bit k bit n bit

 s (sign) sepanjang satu bit mengkodekan bit tanda s

 exp (exponent) sepanjang k bit mengkodekan eksponen E

 frac (fraction) sepanjang n bit mengkodekan signifikan M

Ukuran (tipe data float pada C)

 Single precision : s=1 bit, exp=8 bit , frac=23 bit → total 32 bit

 Double precision : s=1 bit, exp=11 bit, frac=52 bit → total 64 bit