关于浮点数的学习

十进制数转IEE754单精度浮点数

对学习内容的理解

例题

5.75→01000000101110000000000000000000
161.875→01000011001000011110000000000000
-0.0234375→10111100110000000000000000000000

python编写IEEE754程序

def ConvertFixedDecimalToComplement(fixedDecimal) :
fixedpoint = int(fixedDecimal) / (10.0**len(fixedDecimal))
s = ''
while fixedDecimal != 1.0 and len(s) < 23 :
fixedpoint = fixedpoint * 2.0
s += str(fixedpoint)[0]
fixedpoint = fixedpoint if str(fixedpoint)[0] == '0' else fixedpoint - 1.0
return s

def ConvertToExponentMarker(number) : 
return bin(number + 127)[2:].zfill(8)

def ConvertToFloat(floatingPoint) :
floatingPointString = str(floatingPoint)
if floatingPointString.find('-') != -1 :
sign = '1'
floatingPointString = floatingPointString[1:]
else :
sign = '0'
l = floatingPointString.split('.')
front = ConvertFixedIntegerToComplement(int(l[0]))
rear = ConvertFixedDecimalToComplement(l[1])
floatingPointString = front + '.' + rear 
relativePos = floatingPointString.find('.') - floatingPointString.find('1')
if relativePos > 0 :
exponet = ConvertToExponentMarker(relativePos-1)
mantissa = floatingPointString[floatingPointString.find('1')+1 : floatingPointString.find('.')] + floatingPointString[floatingPointString.find('.') + 1 :]
else :
exponet = ConvertToExponentMarker(relativePos)
mantissa = floatingPointString[floatingPointString.find('1') + 1: ]
mantissa = mantissa[:23] + '0' * (23 - len(mantissa))
floatingPointString = '0b' + sign + exponet + mantissa
print(floatingPointString)
return hex( int( floatingPointString , 2 ) )

参考

https://blog.csdn.net/qq_40890756/article/details/83111431