Tensor Memory Accelerator
The Tensor Memory Accelerator is an asynchronous data-movement facility introduced with NVIDIA Hopper. It moves tensor regions between global and shared memory without routing every element through a thread register.
How it works
A descriptor describes the tensor layout. A producer issues a transfer, while consumer warps can perform independent work. A completion barrier tells consumers when the destination data is ready. Correct synchronization still governs when a buffer can be read or reused.
Work through an example
A tiled matrix kernel can compute on shared-memory buffer A while TMA fills buffer B. Swapping their roles is safe only after the transfer and the previous readers have completed.
Where the term can mislead
TMA is a transfer engine. Tensor Memory on supported Blackwell devices is a separate storage resource; the similar names do not make them interchangeable.
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.