Moving computations whose values do not change across loop iterations OUT of the loop body is called:
Loop-Invariant Code Motion identifies expressions whose operands remain constant across all loop iterations and hoists them into a pre-header outside the loop.
Replicating a loop body multiple times to reduce loop control overhead (branching and counter updates) is called:
Loop Unrolling duplicates the loop body to decrease branch testing frequency and exploit instruction-level parallelism.
Combining two adjacent loops that execute over the same index bounds into a single loop is known as:
Loop Fusion merges two adjacent loops sharing iteration space into one loop, improving data cache locality and eliminating extra loop test overhead.
In Data-Flow Analysis, a variable x is defined as 'Live' at point p if:
Liveness Analysis: A variable is live at point p if its current value will be read downstream along at least one execution path before being overwritten.
Live Variable Analysis is classified as which type of Data-Flow Analysis framework?
Liveness flows backwards (from uses to definitions) and uses union (U) path combination (May-analysis: live if used along AT LEAST ONE path).