AMD ROCm GPU Computing Stack | Generated by AI
This is the ROCm monorepo — AMD’s open-source GPU computing platform.
It’s a meta-repository (the official ROCm/ROCm repo on GitHub) that documents and organizes the entire ROCm ecosystem, which includes:
- HIP — C++ runtime API & kernel language (CUDA-compatible)
- Math libraries — rocBLAS, rocFFT, rocRAND, rocSPARSE, etc.
- Runtimes & compilers — ROCclr, ROCm-CompilerSupport, OpenMP, OpenCL
- Profiling & debugging — rocprofiler, roctracer, ROCdbgapi
- Deep learning — integrations with PyTorch, TensorFlow
- Communication libraries — RCCL (like NCCL for AMD)
Think of it as the AMD equivalent of NVIDIA’s CUDA ecosystem — a full stack for GPU-accelerated HPC, AI/ML, and scientific computing on AMD GPUs (Instinct, Radeon). This particular repo is the top-level manifest/doc repo that ties all the sub-projects together.
Generated by AI. Curating and sharing still takes effort.
If you find it useful, feel free to
donate.
WeChat: @lzwjava
·
X: @lzwjava ·
Say hi 👋