The CPU
The general-purpose brain that runs everything around the GPUs.
How big?About the size of a large cracker.
Real: What it looks like.
Keys: arrows rotate · + / − zoom · 0 reset · 1–4 views · S signal · T tour · L labels · Space spin. Models are stylised and built from code: proportions are honest, details are simplified.
What it is
A CPU has fewer cores than a GPU, but each one is powerful and clever. It predicts branches, reorders work, and handles any kind of program.
Big server CPUs are built from chiplets: several core dies around a central I/O die, all under a metal heat spreader. They sit in a socket on the motherboard.
Why AI needs it
In an AI server the CPU is the coordinator. It runs the operating system, loads and shuffles training data, launches work on the GPUs and talks to the network. A slow CPU can leave expensive GPUs waiting.
Every labeled part
- 1
Heat spreader
The metal lid you normally see. X-ray or explode to look underneath.
- 2
Core chiplet
Each small die holds several cores. Signal mode lights them as work is scheduled.
- 3
I/O die
Connects the chiplets to memory, PCIe (GPUs, NICs, SSDs) and each other.
- 4
Substrate
Green fiberglass-and-copper board that fans the chip's connections out to the pads.
- 5
Contact pads
Thousands of gold pads touch spring pins in the socket.
- 6
Socket
Lets the CPU be installed or replaced without soldering.
Try it · concept lab
CPU vs. GPU: the race
Same 256 multiplications, two very different chips.
CPU · 8 fast cores
0/256 multiplications · 0 ticks
GPU · 256 slower cores
0/256 · done in 4 ticks
Each GPU core is 4× slower here, and it still wins 8× because it does them all at once. Real chips differ in the details; the shape of the result is right.
Big idea: AI math is lots of small, independent pieces. Many slow workers beat a few fast ones when the jobs don't depend on each other.
Swap it: other ways to do the same job
x86
The long-time server standard (Intel, AMD). Huge software compatibility.
Arm
Increasingly common in data centers. Often more power-efficient; used in several CPU+GPU superchips.
GPU
Thousands of simple cores: much faster for AI math, poor at the operating system's branchy, one-thing-at-a-time work.
Talk about it
- Q1
The CPU is the coordinator that keeps the GPUs busy. Who plays that role on a sports team or in a kitchen?
- Q2
A slow coordinator can leave expensive helpers waiting around. Have you ever been stuck waiting on one person? What fixed it?
For grown-ups: there are no right answers here. Ask a question, then ask "why do you think that?" The reasons matter more than the answer.
Printable question sheet (PDF)