GPU Programming

CUDA C++.

GPU kernel development with CUDA C++ — thread hierarchies, shared memory, warp primitives, and performance tuning.

Beginner

Start here — no prior experience needed

1 lesson

Learn kernel execution, thread/block indexing, memory access qualifiers, and write a basic CUDA C++ kernel.

For developers with core concepts down

1 lesson

Use shared memory for tiling, avoid bank conflicts, and apply warp-level primitives for fast intra-warp reductions.

Production-grade patterns for experienced engineers

1 lesson

Use atomics safely, build efficient parallel reductions, and understand performance/contentsion trade-offs in CUDA C++.