Vector-length agnostic (VLA) refers to a programming model whereby vectorized code is compiled such that it could run under any vector length the underlying microarchitecture implements. In other words, VLA-based vectorized code automatically adapts and scales to the vector length of the machine at runtime.

This article is still a stub and needs your attention. You can help improve this article by editing this page and adding the missing information.