Processing math: 100%
Skip to content

15-213/14-513/15-513: Introduction to Computer Systems

约 362 个字 11 行代码 6 张图片 预计阅读时间 2 分钟 共被读过 4 次

1st Lecture, Aug 27, 2024

📌 Instructors

  • 15-213/15-513:
    🧑🏫 Phil Gibbons
    🧑💻 Brian Railing
  • 14-513:
    🧑🏫 Mohamed Farag
    🧑💻 David Varodayan

🎯 Course Theme: Systems Knowledge is Power!

Key Outcomes:

  1. Become a better programmer:
    - Find and eliminate bugs efficiently.
    - Understand and tune program performance.
  2. Prepare for advanced systems courses:
    - Compilers, OS, Networks, Computer Architecture, Embedded Systems, etc.

🌟 The Five Great Realities

Great Reality #1: Ints ≠ Integers, Floats ≠ Reals

Example 1: Is x2≥0?

  • Floats: Always true.
  • Ints:
  • 40000×40000=1600000000 ✅
  • 50000×50000= Overflow ❌ (Result: −2147483648 for 32-bit ints).

Example 2: Is (x+y)+z=x+(y+z)?

  • Ints: ✅ (Associative).
  • Floats: ❌
  • (1e20+−1e20)+3.14=3.14
  • 1e20+(−1e20+3.14)=0 (Catastrophic cancellation due to limited precision).

Great Reality #2: You Must Know Assembly

  • Why?
  • Debugging: High-level abstractions break down.
  • Performance tuning: Understand compiler optimizations.
  • System software: Compilers/OS rely on assembly.
  • Focus: x86 assembly.

Great Reality #3: Memory Matters

Memory Referencing Bug Example:

C
typedef struct {  
    int a;  
    double d;  
} struct_t;  

double fun(int i) {  
    volatile struct_t s;  
    s.d = 3.14;  
    s.a[i] = 1073741824;  // Out-of-bounds access  
    return s.d;  
}

Results:
- fun(0) → 3.14 ✅
- fun(3) → 2.00000061035156 ❌ (Corrupted memory).

Great Reality #4: Performance ≠ Asymptotic Complexity

  • Constant factors matter!
  • Example: Loop unrolling, cache locality.
  • Memory Hierarchy:
  • L1/L2 cache vs. RAM vs. Disk (4.3ms vs. 81.8ms access time).

Great Reality #5: Computers Do More Than Execute Programs

  • I/O and Networking:
  • Concurrency, unreliable media, cross-platform issues.
  • Example: Web proxy lab (L7) → Handle HTTP requests concurrently.

📚 Course Components

Labs (50% of Grade)

  1. Data Lab: Bit manipulation.
  2. Bomb Lab: Reverse engineering.
  3. Attack Lab: Code injection.
  4. Cache Lab: Optimize memory locality.
  5. Malloc Lab: Implement malloc/free.
  6. Shell Lab: Build a Unix shell.

Written Assignments (20%)

  • Peer-reviewed problem sets (drop lowest 2).

Final Exam (30%)

  • Covers all course concepts.

⚠️ Academic Integrity

Cheating Examples:

  • ❌ Copying code from peers/web.
  • ❌ Using AI tools (ChatGPT, Copilot) for solutions.
  • ❌ Reusing old code.

Consequences:

  • Failing grade, expulsion, or retroactive penalties.

Allowed:

  • ✅ Discuss high-level design.
  • ✅ Use textbook/CS:APP code (with attribution).

💻 Lab Policies

  • Grace Days: 5 per semester (max 2 per lab).
  • Late Penalty: 15% per day after grace days.

📖 Textbooks

  1. Primary: Computer Systems: A Programmer’s Perspective (Bryant & O’Hallaron).
  2. Recommended: The C Programming Language (K&R).

🛠️ Tools & Infrastructure

  • Shark Machines: ssh shark.ics.cs.cmu.edu.
  • Autolab: Submit labs, view scoreboards.

🚀 Key Advice

  • Start labs early!
  • Commit code frequently (Git history matters).
  • Attend bootcamps (GDB, Makefiles, C debugging).