Warp
A warp is a group of 32 NVIDIA CUDA threads through which the hardware issues instructions. An instruction may act on fewer than 32 lanes when some threads are inactive.
How it works
Threads in a block are grouped into warps using their linearized thread indices. Threads that take different paths can require instructions under different active masks. Warp-level communication must use the correct participation mask and synchronization rules.
Work through an example
A 100-thread block needs four warps. The fourth contains four valid threads and 28 inactive positions, so a multiple-of-32 block size often avoids a partially populated warp.
Where the term can mislead
Warp size is not a portable synonym for an AMD wavefront size. Check the architecture and programming API before transferring warp-level code.
Connect the idea
Take it into practice
Continue with the existing concept, apply it to an interview problem, or follow a course lesson. Access requirements are shown before you open a page.
New to GPUs? Start with the free execution-model concept.
Sources and scope
Examples here are teaching calculations, not measured GPU performance. Consult the linked documentation for the exact device, instruction and tool version. Source review: 25 October 2026.