CSU, Chico: Computer Science 380  (Computer Architecture)    

Enlarge | Normal Size | Reduce Slide of 22 (Refresh off)

Ch 4: EXPLOITING INSTRUCTION-LEVEL PARALLELISM THROUGH SOFTWARE APPROACHES Basic pipeline scheduling Ex 1: Loop unrolling 1/2 Ex 1: Loop unrolling 2/2 Ex 2: Loop unrolling with rescheduling Loop unrolling - defined Ex 3: with 4-fold loop unroll Ex 4: with unrolling and scheduling Ex 5: Eliminating data dependences Ex 6: Eliminating name dependences Limits to loop unrolling Loop unrolling and scheduling with static multiple issue Static branch prediction Static branch prediction helps static scheduling Some static branch prediction schemes Profile-based predictor statictics Predict-taken vs. profile-based More on VLIW ... VLIW processor schematic More on VLIW ... Difficulties with early VLIW Compiler support for ILP This E-Slideshow was prepared by Dr. J for CSCI 380 (Last revised: Thu Jan 30 21:34:39 PDT 2003)