Hi! I have some difficulties understanding how I'm supposed to calculate this 16bit float to base10.
This is something that is coming up on a test so I would be pleased to learn how this is supposed to be done.
The example we got had number 0101010111110000 where first from left is sign bit, next 6bits is exponent and the rest 9bits is mantissa.
The expression is $(-1)^s * 2^{e-b} * 1.m$
I've spent some time reading wikipedia and such, but I just don't get this. I'm very confused since the only examples I bumped into was [1s|5e|10m] (maybe I'm thinking this wrongly in some way or I'm just plainly stupid)
I would also be pleased if I could get an explanation how I transfer base10 decimal like 118.625 to binary float point and that to 16bit float as before [1 sign | 6 exp | 9 mantissa]
Sorry for the inconvenience
/me bows humbly
The raw fields in your number are:
$s = 0$
$e = 101010_2 = 42_{10}$
$m = 111110000_2$
Without knowing the specific format used, we can't be certain what your bit string represents. But if the IEEE conventions are followed, the exponent bias is 31, and the mantissa has an implied leading 1. In this case, the adjusted values are:
$e' = 42_{10}-31_{10} = 11_{10}$
$m' = 1111110000_2$
So the value is $1.11111_2 * 2^{11_{10}} = \dfrac{63_{10}}{32_{10}} * 2^{11_{10}} = 4032_{10}$