Introduction of ADDADDA-CXX: automatic C99 to C++ conversion for improved development and CUDA integration

1 view

Skip to first unread message

Michel GROSS

unread,

Feb 16, 2026, 4:13:55 AM (3 days ago) Feb 16

to ADDA questions and answers

Dear colleagues,

I would like to inform you about the creation of a new fork of ADDA called ADDA-CXX.

This fork provides a Python script that automatically converts the original ADDA codebase from C99 to C++. The primary objective is to enable developers to benefit from a broader ecosystem of modern development tools and workflows, both on Linux and Windows.

The converted code can be compiled using standard toolchains such as gcc and g++, with build systems based on Makefile or CMake. In addition, it can be compiled directly with Visual Studio on Windows, allowing native integration with the MSVC toolchain and its debugging and profiling tools.

This transition to C++ also significantly facilitates future CUDA development. The NVIDIA NVCC compiler handles the CPU portion of CUDA projects using a C++ compiler, namely g++ on Linux and MSVC on Windows. Having the ADDA codebase in C++ therefore simplifies integration with CUDA and reduces compatibility friction during GPU porting.

The purpose of this fork is not to modify the numerical algorithms or scientific functionality of ADDA, but to provide a technically modernized codebase that improves portability, tooling compatibility, and readiness for GPU acceleration.

I hope this initiative will help make ADDA development more accessible and efficient for contributors across platforms.