Snižování latence násobičky s použitím pipeline s proměnnou latencí. Pipeline s proměnnou latencí kombinuje metody Instruction(Value) reuse a Trivial Computation Bypassing. Prezentovaná myšlenka je testována s použitím simulátoru SimpleScalar LLC a zkušebních úloh SPEC2000. (cs)
Using Variable Latency Pipeline, which combines Instruction(Value) reuse and the Trivial Computation Bypassing, for multiplier execution latecency reduction. Presented idea is tested using SimpleScalar LLC simulator and SPEC2000 benchmarks.
Using Variable Latency Pipeline, which combines Instruction(Value) reuse and the Trivial Computation Bypassing, for multiplier execution latecency reduction. Presented idea is tested using SimpleScalar LLC simulator and SPEC2000 benchmarks. (en)