a$.$ The speedup of a parallelized code can be calculated using Amdahl's law, which is given by:

where P is the proportion of the program that can be made parallel $($in this case, $75\%$ or $0\mathrm{.}75),$ and N is the number of processors $($in this case, $4).$ Substituting these values into the equation, we get:

This means that the single core would need to be $2\mathrm{.}67$ times faster to execute at the same speed as the four$-$way parallelized code.

b$.$ The dynamic energy consumed by a processor is proportional to the square of its voltage and its frequency. If we assume that the voltage scales linearly with the frequency, then the energy consumed by the single core running at $2\mathrm{.}67$ times the frequency would be:

The energy consumed by the four cores running at the original frequency would be:

Therefore, the reduction in dynamic energy from using frequency and voltage scaling would be:

This means that using a single core at a higher frequency would actually consume more energy than using four cores at the original frequency.