Definition of “Dynamic Parallelism” :
« Back to Glossary Index
“Dynamic Parallelism” refers to a feature in GPU computing where kernels, or mini-programs, can be created and executed on the GPU without interaction from the CPU. This allows for adaptive, task-based parallelism directly on the GPU, enhancing efficiency and performance.





