AMD ROCm AI ecosystem

The AMD ROCm AI ecosystem documentation portal includes guides covering deep framework installation and setup, large-scale model training, LLM and diffusion inference serving, and AI workload performance optimization on AMD GPUs. The ROCm AI ecosystem lives on top of the ROCm Core SDK, which provides the underlying GPU runtimes (HIP), compilers, and math libraries.

Deep learning frameworks

Install PyTorch and JAX on AMD GPUs. Includes hardware-specific instructions for AMD Instinct and Radeon GPUs and Ryzen APUs across Linux and Windows using pip.

• Install PyTorch

• Install JAX

Training

Scale model training across multiple AMD GPUs using PyTorch distributed primitives (DDP, RPC, collective communication) for large models that exceed single-GPU memory.

• Scale model training

Inference

Serve LLMs and generative AI models using high-performance inference frameworks. Covers single-node and distributed multi-GPU deployments.

• vLLM

• SGLang

• MIGraphX

• ONNX Runtime

• xDiT

• ComfyUI

Distributed inference

Multi-node prefill-decode disaggregated serving over RDMA networking using MoRI (Modular RDMA Interface) on MI355X clusters.

• vLLM with MoRI recipe

• SGLang with MoRI recipe

Optimization

Improve throughput, latency, and memory efficiency for AI workloads on AMD Instinct GPUs.

• Workload optimization

• vLLM V1 performance

• Model quantization

• Model acceleration libraries

• Triton kernels

• Composable Kernel

Tutorials

Hands-on guides and recipes for building AI applications on AMD hardware.

• AI Playbooks

• AI Developer Hub