Warp scheduler
A warp scheduler chooses an instruction from a ready resident warp and issues it to an available execution pipeline.
How it works
Residency gives a warp a place on the SM. Eligibility means its next instruction can proceed under the current dependency and resource conditions. A scheduler can have many resident warps but no eligible work when all are waiting.
Four resident warps. Two ready. One selected.
Work through an example
In an illustrative cycle, four warps are resident, two are eligible and one is selected to issue. The other eligible warp is ready but not selected; the remaining two are waiting.
Where the term can mislead
Adding resident warps helps only if it supplies useful independent work. It cannot remove a dependency inside one thread.
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.