Simplified 8-Bit Floating-Point Multiplier

Simplified 8-bit floating-point multiplier in 1+3+4 format (1 sign bit, 3 biased exponent bits, 4 mantissa bits + 1 implicit).

Algorithm (combinational):

1. Sign: sign_r = sign_a XOR sign_b.

2. Exponent: exp_r = exp_a + exp_b - bias (bias = 3).

3. Mantissa: mant_r = mant_a × mant_b, including the implicit 1 of both operands.

4. Normalisation: if the multiplication overflows (mant_r ≥ 2), right-shift and increment exp_r.

5. Truncation: keep the top 4 bits of the normalised mantissa.

*Special cases (NaN, infinity, subnormals) are not handled: educational version.*