Problems That Love the GPU
Spotting embarrassingly parallel workloads.
Not Every Job Fits the GPU
GPUs are powerful, but they shine on a specific shape of problem. Learning to spot that shape saves you a lot of wasted effort. 🔍
Embarrassingly Parallel Work
The dream case is embarrassingly parallel work: many items, each handled the same way, with no item depending on another.
All lessons in this course
- CPU vs GPU: Latency vs Throughput
- SIMT: The Same Instruction, Many Threads
- What CUDA Actually Is
- Problems That Love the GPU