Source: rocm Section: devel Standards-Version: 4.7.3 Vcs-Git: https://salsa.debian.org/rocm-team/rocm.git Vcs-Browser: https://salsa.debian.org/rocm-team/rocm Maintainer: Debian ROCm Team Uploaders: Cordell Bloor , Christian Bayle , Build-Depends: debhelper-compat (= 13), sphinx-common, Build-Depends-Indep: dh-sequence-sphinxdoc , python3-rocm-docs (>= 1.23.0-1~) , python-rocm-docs-doc , python3-sphinx-reredirects , python3-sphinx-sitemap , dh-sequence-python3, pybuild-plugin-pyproject, python3-all, python3-sphinxcontrib.datatemplates , libjs-underscore , libjs-jquery , python3-doc , Testsuite: autopkgtest-pkg-pybuild Homepage: https://github.com/ROCm/ROCm # Needs further discussion. Kitchen sink package seems useful. # #Package: rocm #Architecture: amd64 arm64 ppc64el #Depends: rocm-dev, # rocm-tests, # rocm-doc #Description: All ROCm components # This metapackage comprises all ROCm components available, including # development tools, libraries, tests and documentation. Package: rocm-dev Architecture: amd64 arm64 ppc64el Depends: ${misc:Depends}, amd-dbgapi-dev, hip-utils, hipcc, hipify-perl, libamd-smi-dev [amd64], libamdhip64-dev, libhipblas-common-dev, libhipblas-dev [amd64 arm64], libhipcub-dev, libhipfft-dev [amd64 arm64], libhiprand-dev, libhipsolver-dev [amd64 arm64], libhipsparse-dev, libmiopen-dev [amd64 arm64], liboam-dev, # librccl-dev, # broken on unstable librocalution-dev, librocblas-dev [amd64 arm64], librocfft-dev [amd64 arm64], librocm-smi-dev, librocprim-dev, librocrand-dev, librocsolver-dev [amd64 arm64], librocsparse-dev, librocthrust-dev, libroctx-dev, rocminfo, Recommends: librpp-hip-dev [amd64], Suggests: liboffload-21-dev, libomp-21-dev, rocm-opencl-icd, Description: Tools and libraries for ROCm development This metapackage provides development tools, libraries, and headers used when developing for the AMD ROCm platform. The development tools include a HIP compiler, and the libraries include a wide range of math, machine learning, inter- and intra-node communication, and basic primatives for authoring programs with AMD GPU acceleration. Package: rocm-tests Architecture: amd64 arm64 ppc64el Depends: ${misc:Depends}, ${perl:Depends}, # libhipblas2-tests, # todo libhipcub-tests, libhipfft0-tests [amd64 arm64], libhiprand1-tests, libhipsolver0-tests [amd64 arm64], libhipsparse1-tests, libmiopen1-tests [amd64 arm64], # librccl1-tests, # broken on unstable librocblas4-tests [amd64 arm64], librocfft0-tests [amd64 arm64], librocprim-tests, librocrand1-tests, librocsolver0-tests [amd64 arm64], librocsparse1-tests, librocthrust-tests, Description: Tests for validating the ROCm software stack This package contains test utilities that can be used to verify that the ROCm software stack is functioning correctly. Package: rocm-doc Section: doc Architecture: all Multi-Arch: foreign Build-Profiles: Depends: ${sphinxdoc:Depends}, ${misc:Depends}, Description: Documentation for the AMD ROCm software stack ROCm is an open-source stack primarily composed of open-source software for performing compuations on graphics processing units (GPUs). ROCm comprises drivers, development tools, and libraries that enable GPU programming from low-level kernel integrations to high-level applications. . With ROCm you can customize your GPU software to meet your specific needs. You can develop, collaborate, test, and deploy your applications in a free, open-source, and secure software ecosystem. ROCm is well-suited for high-performance computing (HPC), artificial intelligence (AI), scientific computing, and computer aided design (CAD). . ROCm is powered by the Heterogeneous-computing Interface for Portability (HIP). HIP is an open-source software C++ GPU programming language and its corresponding runtime. It allows developers to create portable applications by deploying code for a wide range of hardware, from desktop gaming GPUs to exascale HPC server clusters. . ROCm supports programming models such as OpenMP and OpenCL. It includes all necessary open source software compilers, debuggers, and libraries. ROCm support is integrated into machine learning (ML) frameworks, such as PyTorch and TensorFlow. . This is the common documentation package.