IEEE 754 Floating Point Numbers Flashcards
what does normalised mean
one number before the radox point
what would a binary fraction look like after removing the hidden bit
.101001`(whatever the fraction was)
what is the fixed bias for IEE754 numbers
127
do you add or subtract the exponent from the bias
add
if the sign bit is set to 1 this means that the fraction is
negative
if the sign bit is set to 0 this means that the fraction is
positive
if all the exponent and fraction bits are set to 0, what number is being represented
0
if the exponent bits are all set to 1 and the fraction bits are all set to 0 what number is being represented
infinity
if the exponent bits are all set to 1 and the fraction bits are anything but all 0, what number is being represented
not a number
what is the bias for double precision, 64 bit ieee 754 numbers
1023
is 32 bit precision sufficient
yes, more than enough for the decisions needed in AI and machine learning
to add two floating points together, what needs to be the same
the exponents
if the two exponents are different when adding floating point numbers, how do you change this so that theyre equal
shift the smaller one to the right until its exponent is the same as the larger one
after a decimal point what does the power become
minus numbers
how to convert from decimal to binary
do the whole number part as normal
multiply the fraction part by 2 continuously until there is nothing after the decimal point ie a whole number, read the whole parts downwards
