Modern Computer Architecture
Why chips got fast (pipelines, caches, and parallelism) and why your code's speed depends on the silicon.
AI tutor is turned off for this class
Use the CS80E lectures, notes, and assignments below to keep learning.
Lesson overview
Why chips got fast (pipelines, caches, and parallelism) and why your code's speed depends on the silicon.
Teacher script · 45 min
- 0–5
Hook
A processor doesn't do one whole instruction then the next; like a car factory, it splits work into stages and keeps all stages busy on different instructions at once. This 'pipelining' is a big reason chips feel fast: many things half-done simultaneously beats one thing done fully at a time.
- 5–15
Explore
Students do the activity in pairs: Cook from a recipe keeping the 5 ingredients you use most on the counter, the rest in the pantry. Fewer pantry trips = faster.
- 15–30
Explain
Clock speeds stopped climbing, so chips went wide: multiple cores, vector units doing one operation on many numbers at once, and GPUs with thousands of tiny cores. Modern speed comes from parallelism, which is exactly why training neural nets lives on GPUs. The catch: not all work splits cleanly.
- 30–40
Connect to the summit
Read aloud the "From the summit" section below. Ask students which ideas from today they recognize in it.
- 40–45
Check
Run the formative check below. Anyone who can explain a key term in their own words has it.
Student activity
Cook from a recipe keeping the 5 ingredients you use most on the counter, the rest in the pantry. Fewer pantry trips = faster.
Slides
Formative check
- 1.In your own words, what is "Pipelining"? (Looking for: Overlapping the stages of many instructions like an assembly line.)
- 2.In your own words, what is "Cache"? (Looking for: A small, very fast memory holding recently-used data near the processor.)
- 3.In your own words, what is "Memory hierarchy"? (Looking for: Layers of storage trading speed for size, from registers to disk.)
Carry-away concepts
- Pipelining
- Overlapping the stages of many instructions like an assembly line.
- Cache
- A small, very fast memory holding recently-used data near the processor.
- Memory hierarchy
- Layers of storage trading speed for size, from registers to disk.
- Parallelism
- Doing many computations simultaneously across cores or units.
From the summit · the Stanford source
You learn how processors execute instructions efficiently: pipelining, memory hierarchy, caching, and parallelism.
This module is based on a Stanford course.