intermediate · gpu
Why GPUs Win Some Workloads and Lose Others
Why are GPUs dramatically faster than CPUs for some workloads and slower for others?
Practical scenario
A team proposes moving an image filter to the GPU. It runs on 200×200 thumbnails, one at a time, as part of a request path.
What it tests
- • Understanding of throughput versus latency design
- • Awareness that transfer cost can dominate
- • Knowledge of divergence and its consequences