Installer adaptivecpp avec Homebrew, Nix, zypper

Liens du projet

• Page d'accueil
• Dépôt
• Docs amont
• Page du gestionnaire de paquets

Commandes et alias

• acpp
• acpp-appdb-tool
• acpp-hcf-tool
• acpp-info
• acpp-pcuda-pp
• syclcc
• syclcc-clang

Historique du projet

The README states that AdaptiveCpp was formerly known as hipSYCL and Open SYCL. It presents the project as a modern platform for C++ heterogeneous programming models that targets CPUs and GPUs from major vendors.

Development is primarily led by Heidelberg University with community contributions. The README describes AdaptiveCpp as both a production platform and a research project, with publications covering SYCL beyond OpenCL, single-pass compilation, standard parallelism offloading, CPU compiler techniques, and JIT-time optimization.

Historique d'adoption

AdaptiveCpp's adoption story is tied to cross-vendor GPU computing. The README says it can target Intel, NVIDIA, AMD, Apple, x86, ARM, and other LLVM-supported CPUs, and that it is used in production on machines of all scales, including some of the most powerful supercomputers.

The release history shows the project moving from hipSYCL 0.x releases to AdaptiveCpp date-based releases beginning with the 23.10 line.

Modes d'utilisation

Users compile applications with `acpp` as a compiler-like frontend, select targets through command-line arguments or environment variables, and integrate AdaptiveCpp into CMake projects through the official usage documentation.

Pourquoi les passionnés de paquets s'y intéressent

AdaptiveCpp is packaging-heavy because it sits at the intersection of LLVM, GPU runtimes, vendor toolchains, and C++ build systems. A package needs to expose multiple executables while preserving compiler discovery, backend support, and environment-variable behavior.

Chronologie

• 2018: GitHub repository now known as `AdaptiveCpp/AdaptiveCpp` was created.
• 2019: hipSYCL 0.8.0 release was published.
• 2020: hipSYCL 0.9.0 release was published and the project cited SYCL beyond OpenCL work.
• 2023: AdaptiveCpp 23.10.0 alpha and 23.10.0 releases marked the renamed release line.
• 2025: AdaptiveCpp 25.x releases continued the date-based release series.

Related projects

• Related projects and technologies include SYCL, OpenCL, CUDA, HIP, LLVM, C++ standard parallelism, oneAPI, and vendor GPU tooling from Intel, NVIDIA, AMD, and Apple.

Sources

• https://api.github.com/repos/AdaptiveCpp/AdaptiveCpp
• https://api.github.com/repos/AdaptiveCpp/AdaptiveCpp/releases
• https://github.com/AdaptiveCpp/AdaptiveCpp#readme