What is CUDA? 10 XP Easy

Ctrl+Enter Run Ctrl+S Save

Welcome to The Parallel Universe Chronicles! ⚡ You are about to become a commander of thousands of tiny, blazing-fast soldiers called GPU threads.

💡 Imagine this: You need to paint a giant wall. You could do it alone (that's a CPU — one strong worker). Or you could hire 10,000 tiny painters, each painting one small section at the same time (that's a GPU!). The GPU finishes in seconds what takes the CPU hours.

CUDA stands for Compute Unified Device Architecture. It's a platform and programming model created by NVIDIA in 2007 that lets you write programs that run on the GPU.

🎮 GPU = Graphics Processing Unit — originally made for games
⚡ CUDA = The language/framework to program NVIDIA GPUs
🤖 Why it matters — ChatGPT, Stable Diffusion, Tesla Autopilot all run on CUDA
🚀 Speed — A modern GPU has 10,000+ cores vs 16-32 CPU cores
🏆 Who uses it — Google, Meta, Tesla, NVIDIA, every AI company

CUDA programs are written in C/C++ with special CUDA extensions. The code you write runs on TWO devices: the Host (your CPU) and the Device (your GPU). They work together like a general (CPU) commanding an army (GPU).

// A CUDA program lives in two worlds:

// HOST CODE (runs on CPU)
int main() {
    // 1. Prepare data on CPU
    // 2. Send data to GPU
    // 3. Tell GPU what to do (launch kernel)
    // 4. Get results back from GPU
    return 0;
}

// DEVICE CODE (runs on GPU — called a "kernel")
__global__ void myKernel() {
    // This runs on THOUSANDS of threads simultaneously!
}

🌍 Real-world CUDA applications:

🧠 AI & Deep Learning — Training neural networks (PyTorch, TensorFlow use CUDA under the hood)
🎬 Video rendering — Pixar & Disney render movies using GPU farms
🔬 Scientific simulation — Protein folding, climate modeling, physics simulations
💊 Drug discovery — Simulating drug molecule behavior
🔐 Cryptography — Bitcoin mining, encryption cracking

🎯 Your journey begins now. By the end of this course, you'll understand how the world's most powerful computing systems work — and you'll be able to build them yourself!

📋 Instructions

Let's make sure your CUDA environment is set up and you understand the basics! Write a program that prints an intro to CUDA: ``` Welcome to CUDA Programming! GPU has: 10496 CUDA cores CPU has: 16 cores Speedup possible: 656x ``` (Use any numbers you like — the key is to print 4 lines with those labels.)

This is just C code — use printf() to print each line. Example: printf("Welcome to CUDA Programming!\n"); The \n creates a new line.

⚠️ Try solving it yourself first — you'll learn more!

#include <stdio.h>

int main() {
    printf("Welcome to CUDA Programming!\n");
    printf("GPU has: 10496 CUDA cores\n");
    printf("CPU has: 16 cores\n");
    printf("Speedup possible: 656x\n");
    return 0;
}

← Back to Course Next Exercise →

main.py

Hi! I'm Rex 👋

Output

Ready. Press ▶ Run or Ctrl+Enter.

›