Search

super repo for rocm libraries

High-performance LLM, VLM, embedding, reranking, and audio inference for Apple Silicon, NVIDIA CUDA and AMD ROCm systems.

Modular RDMA Interface

HIP: C++ Heterogeneous-Compute Interface for Portability

This repository is a fork of the upstream AMDGPU Driver, used by the ROCm project. To submit a patch, please follow the guidelines at https://docs.kernel.org/process/submitting-patches.html#submittingpatches . For reporting an issue, please open a Bug Report at https://gitlab.freedesktop.org/drm/amd/-/issues

AMD ROCm™ Software - GitHub Home

Backend.AI is a streamlined, container-based computing cluster platform that hosts popular computing/ML frameworks and diverse programming languages, with pluggable heterogeneous accelerator support including CUDA GPU, ROCm GPU, Gaudi NPU, Google TPU, GraphCore IPU and other NPUs.

Kubernetes (k8s) device plugin to enable registration of AMD GPU to a container cluster

AMD's graph optimization engine.

HIPIFY: Convert CUDA to Portable C++ Code

Antares: an automatic engine for multi-platform kernel generation and optimization. Supporting CPU, CUDA, ROCm, DirectX12, GraphCore, SYCL for CPU/GPU, OpenCL for AMD/NVIDIA, Android CPU/GPU backends.

Cross-architecture parallel algorithms for Julia's CPU and GPU backends. Targets multithreaded CPUs, and GPUs via Intel oneAPI, AMD ROCm, Apple Metal, Nvidia CUDA.

AI Tensor Engine for ROCm

AMD RAD's multi-GPU Triton-based framework for seamless multi-GPU programming

super repo for rocm systems projects

Fresh builds of llama.cpp with AMD ROCm™ acceleration

AiTer Optimized Model

FlyDSL is the Python front‑end of the project: Flexible LaYout DSL.

AMD YES!

AI-native IDE with agentic workflows, iPhone emulation on Windows/Linux, PyTorch ML Studio, and ROCm-optimized local AI. Built for security researchers and cross-platform developers.

AMD Strix Halo local LLM guide: setup for Ryzen AI MAX+ 395 / Radeon 8060S, Ollama, llama.cpp Vulkan/RADV, ROCm, raw evidence, direct 100 t/s 30B Qwen, 140 t/s CHADROCK MTP, 120B GGUF.

Local AI text-to-speech with voice cloning and voice design, powered by GGML. C++17 port of OmniVoice (k2-fsa/OmniVoice). 646 languages, 24 kHz mono output, runs on CPU, CUDA, ROCm, Metal, Vulkan.

Local AI text-to-speech with voice cloning and voice design, powered by GGML. C++17 port of Qwen3-TTS (QwenLM/Qwen3-TTS). 10 languages, 24 kHz mono output, runs on CPU, CUDA, ROCm, Metal, Vulkan.

DeepSeek 4 Flash and PRO local inference engine for Metal, CUDA and ROCm

Kyojin: the Yamz inference engine for AMD Strix Halo (ROCm, gfx1151), built on ExLlamaV3. Runs 300B-class MoE models on one 128 GB mini PC.