The CUDA Handbook: A Comprehensive Guide to GPU Programming

Glossary / TLA Decoder

aliasing – Creating more than one way to access the same memory. Examples: A mapped pinned buffer in CUDA is aliased by a host pointer and a device pointer; a texture reference bound to device memory aliases the device memory.

AOS – See array of structures.

API – Application programming interface.

array of structures – Memory layout in which the elements that describe an object are contiguous in memory (as if declared in a structure). Contrast with structure of arrays.

asynchronous – Function calls that return before the requested operation has been performed. For correct results, CUDA applications using asynchronous operations subsequently must perform CPU/GPU synchronization using CUDA streams or events.

computational ...