CS382M: Lectures from Spring 1997

• notes1

• notes2

• Lecture3: Cost, margins, geometric mean notes3

• Lecture4: DLX Pipeline and Instruction Set notes4

• Lecture5: Forwarding to avoid data hazards in DLX pipeline. notes5

• Lecture6: DLX Pipeline for branches. notes6

• Lecture7: Compiler reordering for ILP. notes7

• Lecture8: Case study: Scoreboards (Slides courtesy Randy Katz of UC Berkeley). notes8

• Lecture9: Case study: Tomasulo Algorithm (Slides courtesy Randy Katz of UC Berkeley). notes9

• notes10

• Lecture11: Branch Prediction. notes11

• notes12

• Lecture13: ILP. notes13

• notes14

• Lecture15: Reducing Cache Misses. notes15

• notes16

• Lecture17: Reducing Cache Misses. notes17

• notes18

• Performance of AXP mem system. notes19

• Patterson's IRAM talk. Another talk. IRAM class lecture Patterson's IRAM paper.

• Lecture 21: CPU/Memory Wrap-up. Supplemental reading: Cache Performance, Load Value Prediction

• Lecture 22: I/O Intro and Disks notes22

• Lecture 23: Disks, Response Time, and a Little Queuing Theory Slides notes23

• Lecture 24: Response Time (See lecture 23 slides)

• Lecture 25: A little Queuing theory, Connecting I/O to processors notes25

• Lecture 26: Connecting I/O to processors, benchmarking notes26

• Lecture 27: RAID, Benchmarks notes27

• Lecture 28: RAID, Tapes, Benchmarks notes28

• Lecture 29: Networks Intro notes29

• Lecture 30: Networks: HW and SW Interfaces notes30

• Lecture 31: Networks: Beyond the NI notes31

• Lecture 32: Networks: Beyond the NI notes32

• Lecture 33: Intro to Parallel Architectures notes33

• Lecture 34: Coherency: Snooping notes34

• Lecture 35: Directory-based Coherency; Consistency notes35

• Lecture 36: Consistency, Large-scale MPs notes36

• Lecture 37: Exam postmortem, Project presentation guidelines., Course evaluations.