Compiling method, compiling device, electronic device, storage medium and program product

Abstract

The disclosure provides a compiling method, a compiling device, an electronic device, a storage medium and a program product, which can be used in the fields of source file compilation and deep learning. The method includes: obtaining a source file collection, the source file collection includes at least one type of source file; obtaining a directory collection of an object collection associated with the source file collection; for the first directory in the directory collection, based on the first directory includes the same A first set of objects associated with the first type of source files, combining the source files associated with the first set of objects into a first subset of source files; and compiling the first subset of source files as a whole. By using the above method, the compilation efficiency can be effectively improved and the compilation time can be reduced by reducing the number of compilations for the source files, thereby improving user experience.

Description

technical field [0001] The present disclosure relates to computer technology, and more specifically, to a compilation method, a compilation device, an electronic device, a computer-readable storage medium, and a computer program product, which can be used in the fields of source file compilation and deep learning. Background technique [0002] Deep learning is a key technology in the field of artificial intelligence, and developers usually carry out related work based on the deep learning framework. The basis of deep learning is artificial neural network, and the core of artificial neural network is matrix operation, so the core of a deep learning framework is matrix operation. The deep learning system is a computing-intensive system, which often involves a large number of highly complex matrix calculations. The smooth operation of the system often requires high processing speed. The central processing unit cannot even meet the needs of matrix calculations. To achieve fast ...

AI Technical Summary

Problems solved by technology

However, due to the complex compilation process of CUDA, CUDA compilation in large-scale projects is often time-consuming, which affects R&D efficiency

[0004] However, traditional techniques for compiling source files either need to be compiled separately for each source file, or can only merge parts of the source files and compile them very roughly, and basically cannot handle CUDA source files

Therefore, the number of compilations required for projects using CUDA source files is large, and cannot be applied to large projects with many source files

Images