cuBLAS
cuBLAS is NVIDIA’s library of GPU linear-algebra routines. cuBLASLt offers additional controls for supported matrix multiplications, including layouts, algorithm selection and fused epilogues.
How it works
The library selects or accepts implementations suited to the operand shapes, types and configuration. A caller still needs correct dimensions, strides, leading dimensions and compute types. Workspace limits can change which algorithms are available.
Work through an example
Use a supported library multiplication as a baseline before writing a custom GEMM. Compare the same shapes, precision, output computation and warmup conditions.
Where the term can mislead
A library name does not guarantee the same kernel is selected for every shape or release. Preserve the exact configuration in a benchmark record.
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.