AI Infra Interviews logo
Reading performance · free reference

Warp execution state

A resident warp holds execution state on an SM. An eligible warp has an instruction ready to issue. An issued warp is one selected by the scheduler at that issue opportunity.

Resident warpActive warpEligible warpIssued warpNot selected

How it works

Resident warps can wait for dependencies, barriers or an unavailable pipeline. Eligible but unselected warps supply alternative ready work. These categories must be read at the same scheduler, cycle and measurement scope.

One illustrative scheduler opportunity

Four resident warps. Two ready. One selected.

All four warps remain resident. C and D are eligible at this opportunity, and the scheduler selects C. This teaching snapshot does not specify a device’s total scheduler count or issue width. Occupancy alone cannot distinguish these states.

Work through an example

Imagine four resident warps: A waits for a load, B waits at a barrier, and C and D are ready. Selecting C leaves D eligible but not selected. Residency is four; readiness is two; one warp issues.

Where the term can mislead

“Active” in an occupancy metric usually means resident. It does not mean the warp issued an instruction in that cycle.

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.

  1. NVIDIA Nsight Compute Profiling Guide ↗