119.90) pils lili jkjklk jljjkl kjhjkh jkjhjh hgg 7: Given the below code is executed on the 5-stage MIPS pipeline with branch evaluation in Execution stage, answer the following questions: Instruction 1: addi $t1, $zero, 4 Instruction 2: addi $t2, szero, 200 Instruction 3: addi $t3, Szero, 100 Instruction 4: add $50, Szero, $zero Instruction 5: add $sl. Szero. Szero Instruction 6: add $s2, Szero, $zero Instruction proc_y: beq $sl, $t3, exit Instruction 8: add $s2, $32, $t2 Instruction 9: proc_x: lw $53, 0 ($80) Instruction 10: add $s3, $s3, $ti Instruction 11: sw $s3, 0($80) Instruction 12: add $50, $30, $ti Instruction 13: bne $50, $s2, proc_X Instruction 14: addi $si, Ssl, 1 Instruction 15: i proc_y Instruction 16: exit: add $v0, $30, $zero a) (3 pts.) Fill in the following table by specifying the instruction number from which the RAW dependency relationship exists, considering only the first iteration of the loops: From Inst # To Instr# Operand $s3 11 $s3 13 $50 10 b) (8 pts.) Assume the code runs on a version of the pipeline organization with no hazard detection and no forwarding. Fill in the table to indicate the minimum number of noops the compiler will insert into the code for correct operation after each instruction. If there are no noop insertions, you can write 0. After Instruction 1 2 3 4 Minimum After Minimum After Minimum After Minimum # of Nops Instruction # of Nops Instruction # of Nops Instruction # of Nops 5 9 13 6 10 14 7 11 15 8 12 16 c) (3 pts.) Considering all executed instructions, including the noops compiler inserted, what is the utilization of the memory write-port in this pipeline while executing the given code. Provide your answer to the nearest integer in %, without decimal points. Data forwarding has been added to the pipeline as a feature for all stages that may benefit from it. Assume there is no detection or handling of control hazards in this part. d) (6 pts.) What is the speedup in this processor performance when the code is executed with data forwarding features compared to when it is executed without any data forwarding (provide your answer with 2 decimal places)? e) (5 pts.) For each of the following instructions, select which unit processes the given source operand and which pipeline staging register last sources the specified source operand for the given processing block and instruction. Only consider the first iteration of any loop. Instruction # Source Operand Processing Block Sourcing Pipeline Staging Register Szero 7 $51 $s3 13 $so 15 proc_y 4 11 f) (5 pts.) Calculate the number of cycles that can be reduced compared to part (a) while running the code on this processor if always-taken static branch prediction is implemented with full BTB and BHT features. Assume the BHT and BTB are large enough to accommodate each branch with a different entry. Jumps do not use BTB. g) (5 pts.) How does your answer to (f) change if 2-bit dynamic branch prediction is implemented with below state diagram, and initial state of weakly-predict-taken for each branch occurrence. Not taken Taken Predict Taken Taken Predict Taken Taken Not taken Not taken Predict Not Taken Predict Not Taken Not taken Taken
