Disclaimer
Everything listed here is “fair game” for the exam, but I cannot guarantee that I haven’t forgotten something
here that just has to be included on the exam. If in doubt, ask on the
Course Discussion Board.
Reading Assignments
- All material from Exam 2.
- Figure 4.17: the role played by each element and control signal in the execution of R-type, lw/sw, and beq instructions.
- The rationale behind pipelining
- Compare single-cycle and pipelined performance
- What is a balanced pipeline; what determines the clock speed of a pipelined processor.
- Figure 4.51: the role played by each element and control signal in the execution of R-type, lw/sw, and beq instructions; the control and data information in each pipeline register
- Structural, Data, and Control hazards
- Register forwarding concepts and implementation
- Spatial and temporal locality: what makes the memory hierarchy “work”
- Hit and miss ratios: calculating the effective (average) memory latency (access time)
- Definitions of cache lines, memory blocks; tag, index, block offset, word offset fields of an address
- Compute widths of address fields given cache configuration parameters, and vice-versa
- Set-associative cache design; address field sizes with different set sizes
- Definitions of write-through and write-back
- DRAM compared to SRAM with respect to capacity, latency, and cost per bit; relate to memory hierarchy principles
- Disks: platters, surfaces, read/write heads, tracks, cylinders, sectors
- Disk latency components and how to compute them
- I/O busses and controllers