GPU Glossary
/
Device Software·1 min read

Thread Block Grid

The collection of thread blocks created by a kernel launch, the top level of the CUDA thread hierarchy, spanning the whole GPU.

A thread block grid is the collection of threads created when a CUDA kernel is launched. Grids are made of thread blocks and can be one-, two- or three-dimensional.

Thread block grids are the highest level of the thread group hierarchy of the CUDA programming model (left). They map onto multiple Streaming Multiprocessors (right, bottom). Modified from diagrams in NVIDIA's CUDA Refresher: The CUDA Programming Model and the NVIDIA CUDA C++ Programming Guide.
Thread block grids are the highest level of the thread group hierarchy of the CUDA programming model (left). They map onto multiple Streaming Multiprocessors (right, bottom). Modified from diagrams in NVIDIA's CUDA Refresher: The CUDA Programming Model and the NVIDIA CUDA C++ Programming Guide.

The matching level of the memory hierarchy is global memory.

Thread blocks in a grid are effectively independent units of computation. They execute concurrently, meaning in indeterminate order. That ranges from fully sequential on a GPU with a single Streaming Multiprocessor to fully parallel on a GPU with enough resources to run them all at once.