CUDA quiz

Test your CUDA knowledge with a free interactive quiz — 20 questions with answers and explanations. No signup needed to play.

What is CUDA "unified memory"?

1. What is CUDA "unified memory"?
2. Why are GPUs well-suited to the kind of workloads CUDA targets?
3. Why is CUDA proprietary to NVIDIA hardware, and what is the main cross-vendor alternative for GPU compute?
4. What do the terms "host" and "device" refer to in CUDA terminology?
5. Why do most deep learning practitioners rarely write raw CUDA kernels directly?
6. What is a CUDA "kernel"?
7. What is CUDA?
8. Why is minimizing host-device memory transfer often critical for CUDA performance?
9. What is "warp divergence" and why does it hurt performance?
10. What is a CUDA "stream"?
11. What is the difference between global memory and shared memory on a CUDA GPU?
12. What does cudaMemcpy typically do in a basic CUDA program?

More quizzes