Features Sample Introduction

Overview

Based on Ascend C feature samples, this document introduces features such as Aclnn (GE graph entry) projects, LocalMemoryAllocator, and Barrier separate memory allocation.

Sample List

Directory Name	Description
00_compilation	This sample introduces the implementation of custom operator compilation projects and static Aclnn invocation
01_invocation	This sample introduces the implementation of Aclnn and Aclop operator invocation
02_framework	This sample introduces custom operator implementation methods for PyTorch, TensorFlow, and ONNX frameworks
03_basic_api	This sample introduces the usage of Ascend C Basic API, including data movement, matrix computation, memory vector computation, resource management, synchronization control, system access, atomic operations, and scalar computation
04_aicpu	This sample introduces the implementation of using AI CPU operators for Tiling sink computation
05_tensor_api	This sample demonstrates the implementation of using TENSOR_API to build Matmul operators in multi-core scenarios, with Relu and inline quantization enabled