How to perform floating point signed 2's complement with mantissa of 3 and exponent of 9 for - 7/8

I used dec2bin with however it is fixed point representation. I am using Matlab 2019b

P = dec2bin((mod(num), 2^bits),bits)

[Rik]

This code will not show the binary representation of floats, like double and single. Is that your goal? If so, you can look at the internal function typecast_double_uint64 in ComputeNonCryptHash, which I wrote to do something similar.

[James Tursa]

Signed 2's complement refers to integer formats. How is this related to your desired floating point format?

[Rohit Mallampati]

7/8 goes to 0.875 which I need to convert to floating point 2's complement representation. In fixed point, first we take 7/8 into binary and then add 1 to LSB to get -7/8. However I want to use floating point to represent (fraction in precise).

I can put in in this way, I need to quantise coefficients of a filter to 12 bits (mantissa=3 and exponent= 9) using floating point representation.

[Rohit Mallampati]

The code that I wrote is for fixed point 2's complement.