Question $6$

$1$ pts

One way to achieve parallelism is to have Very Large Instruction Words $($VLIW$).$ Each instruction is actually several bundled together and executed at once using multiple functional units. What is a downside of this approach?

$$ Puts the burden of creating bundles on software $($compilers$)$

Clock speeds must be much slower to accommodate the larger word sizes

Fetch times for the many instructions takes much longer

Each instruction must run one at a time, eliminating the benefit of longer instructions