• 程序
  • a. User mode vs Kernel mode
  • b. 程序的狀態轉換圖

  • 多工協調、死結處理

  • a. 多處理器快取一致性的問題 (Snoopy 快取協調協定)
  • b. Banker's Algorithm
  • c. 多處理器如何協調多程序 , 且可滿足 Mutex, Progress,Bounded Waiting 條件
  • d. 以 Critical Region 解決哲學家晚宴之死結

  • Pipeline CPU

  • a. Hazard
  • b. Branch Prediction Mechnism
  • c. Pipeline Register

  • 快取記憶體架構 (複雜)

  • a. Degree of associativity
  • b. 2-way associative cache system

  • 比較

  • a. Preemptive 與 Non-preemptive scheduling
  • b. Caching 與 Buffering
  • c. Interrupt 與 Trap
  • d. RISC & CISC 的組語語法
  • e. Track free disk space using Free list vs Bit-map

  • 釋義

  • a. Branch prediction mechanism
  • b. System Call
  • c. User / Kernel mode, from Software/Hardware point of view
  • d. Bluck Buffer Cache
  • e. Thrashing
  • f. DMA
  • g. Pipeline Register
  • h. Process-Memory / IO / Backplane BUS

  • 計算

  • a. Effective access time
  • b. 改過的 Pipeline CPU 之增速
  • c. 2-level page table virtual memory system

  • Simulation

  • a. Cache miss/hit
  • b. Virtual addressing
  • c. Page fault
  • d. Gantt Chart

  • Implementation

  • a. Binary Semaphore using Spin-lock

  • 特殊

  • a. 描述 RAID 5 架構,其如何實做讀寫
  • b. Snoopy Cache Coherence Protocol