Theory
The gamer's question
A customer at the shop wants a gaming PC. Chirag quotes two prices: processor ₹18,000, graphics card ₹35,000.
The customer frowns: "Why am I paying more for the graphics card than the actual computer part? Aren't they both processors?"
They are. And they are built on two opposite philosophies. Understanding the split is this topic, one modern exam question, and half of every PC-buying decision.
Theory
One genius vs a thousand clerks
A CPU is a small team of genius professors: each can solve any problem, complex logic, sudden changes of plan, but there are only a few of them. A GPU is a hall of a thousand clerks: each clerk is simple, but give them all the SAME small sum at once and the hall finishes a million sums before the professors clear ten. Different problems need different rooms.
Theory
The two processors, formally
CPU (Central Processing Unit): the general-purpose brain. A few powerful cores (2 to 16 in shop machines) that run the operating system, your programs, and any logic thrown at them, one instruction stream at a time per core.
GPU (Graphics Processing Unit): a specialist. Thousands of simple cores working in parallel. Born to colour millions of screen pixels per frame, the same tiny calculation repeated across huge data, which is why it now also powers video editing and AI.
At a glance
CPU vs GPU at a glance
| Aspect | CPU | GPU |
|---|---|---|
| Cores | Few, powerful | Thousands, simple |
| Built for | Any task, complex logic | Same task on massive data |
| Style | Serial (one after another) | Parallel (all at once) |
| Examples | Running the OS, your C programs | Gaming frames, video render, AI |
Quiz
Which job is the GPU clearly better at than the CPU?
- Applying the same filter to all 20 lakh pixels of a photo
- Running the operating system
- Executing the if-else logic of a billing program
- Reading input from the keyboard
Show the answer
Applying the same filter to all 20 lakh pixels of a photo
Twenty lakh pixels, one identical operation each: that is pure parallel work, the GPU's home ground. The other three are sequential, logic-heavy or coordination jobs, exactly what CPUs are designed for. The pattern to remember: same small task × huge data = GPU; varied logic = CPU.
Think first
Answer the customer
So why does the gaming PC need BOTH an ₹18,000 CPU and a ₹35,000 GPU? Why not just the expensive one? Think about who does what during a game.
Show the answer
The game needs both rooms. The CPU runs the game's logic: enemy decisions, physics, scores, input. The GPU paints every frame: millions of pixels, 60+ times a second. Remove the GPU and frames crawl; remove the CPU and there is no game to draw. They are partners, not rivals, which is why no GPU ever "replaces" a CPU.
Watch out
The mark-losing mistakes
Writing "GPU is faster than CPU". Incomplete, and wrong for most tasks: a GPU is faster only at parallel work; give it branching logic and it loses badly. Second slip: "GPU is only for games". Its parallel muscle now drives AI training and video work, and examiners increasingly expect that modern line. Third: a GPU does not replace the CPU, ever.
Theory
Shop wisdom and what comes next
Buying advice version: office and coding machines live happily on integrated graphics (a small GPU built into the CPU); gaming, editing and AI need a dedicated card with its own memory. Keep the genius-vs-clerks picture: it returns when you meet parallel processing in later semesters, and in every "CPU vs GPU" interview question.
Summary
Key takeaways
- CPU: few powerful general-purpose cores, runs OS and program logic, serial style.
- GPU: thousands of simple cores, same operation on massive data, parallel style.
- GPU wins only at parallelisable work (pixels, video, AI); CPU handles everything else.
- They are partners: every machine needs a CPU; heavy visual/AI work adds a GPU.
- Integrated GPU for office work; dedicated GPU for gaming, editing, AI.
- Memory hook: one genius, a thousand clerks.