Concepts

The Kunpeng Unified Parallel Library (KUPL) provides foundational parallel acceleration functions optimized on the Kunpeng platform, with all interfaces implemented in C/C++ and Assembly. This library provides essential functions, including underlying thread management, task scheduling, thread synchronization, memory allocation, shared memory allocation, shared memory communication, and matrix programming and computing, to maximize the hardware capabilities of Kunpeng processors and provide high-performance basic interfaces.