Wavefront
A wavefront is an AMD GPU group of work-items that execute instructions together. Its width depends on the architecture and execution mode.
How it works
The role resembles an NVIDIA warp, but the names do not establish identical widths, instructions or synchronization semantics. Porting code requires checking wave size, subgroup operations, local data share behavior and compiler support for the target.
Work through an example
A reduction written with a hard-coded 32-lane shuffle tree needs review before moving to a wave64 target. A portable implementation queries or specializes the execution width and validates partial groups.
Where the term can mislead
Similar terminology does not guarantee performance portability. Memory layouts and occupancy limits also differ.
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.