SASS
SASS is the commonly used name for NVIDIA’s native GPU assembly. A cubin stores executable device code and metadata for a particular target.
How it works
Disassemblers show the instructions selected after compiler optimization and register allocation. This is useful for checking whether a kernel uses the expected loads, matrix instructions or spills. Instruction meaning and scheduling details are architecture-specific.
Work through an example
If a source change unexpectedly slows a kernel, compare resource reports and native instructions as well as source code. The compiler may have changed live ranges, instruction count or memory accesses.
Where the term can mislead
A disassembly explains generated work but does not measure its runtime. Combine it with a reproducible timing and profiling experiment.
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.