All Topics
All Topics
Technology
Technology
Design
Design
Programming
Programming
Science
Science
News
News
Gaming
Gaming
Entertainment
Entertainment
Business
Business
Finance
Finance
Sports
Sports
Health
Health
Food
Food
Travel
Travel
Art
Art
Music
Music
Books
Books
Education
Education
Politics
Politics
Personal
Personal
No algorithm. No AI slop. No ads. Just RSS. Pro-human. Indie writers. Real journalism. Open web. Chronological. Hand toasted.

BarraCUDA: Open-Source CUDA Compiler Supports AMD, NVIDIA, and Tenstorrent GPUs

By

rurban

3mo ago· 7 min readenCode

Summary

BarraCUDA is an open-source CUDA C++ compiler written from scratch in C99 that can compile .cu files to multiple GPU architectures including AMD RDNA 2/3/4 binaries, NVIDIA PTX, and Tenstorrent Tensix C++. The project was created as a response to NVIDIA's walled garden approach, aiming to provide a dependency-free compiler without using LLVM. The developer describes the challenge as significant but completed the project anyway, with plans to support more architectures in the future.

Key quotes

· 4 pulled
An open-source CUDA C++ compiler written from scratch in C99 that takes .cu files and compiles them to AMD GPU machine code, NVIDIA PTX, and Tenstorrent Tensix C++, with more architectures planned.
No LLVM, no dependencies, and no permission asked.
This is what happens when you look at NVIDIA's walled garden and think 'how hard can it be?' The answer is: quite hard, actually, but I did it anyway.
Takes CUDA C source code, the same .cu files you'd feed to nvcc, and compiles them to AMD RDNA 2/3/4 binaries, NVIDIA PTX, or Tenstorrent Tensix Metaliu
Snippet from the RSS feed
Open-source CUDA compiler targeting multiple GPU architectures. Compiles .cu to AMD and Tenstorrent GPU's - Zaneham/BarraCUDA

You might also wanna read