This release features:

• AI-assisted GPU management with new ADLX MCP servers. The ADLX MCP servers expose AMD driver features as callable tools for LLM-based agents, enabling AMD GPUs to be controlled and monitored through natural-language requests. See “ADLX MCP Servers” in the ADLX Documentation for the available servers and clients.
• ADLX Python binding (ADLXPybind), built on pybind11 and distributed as an installable Python wheel, providing the foundation the MCP servers use to call ADLX. See “Building Python bindings for ADLX” in the ADLX Documentation.

Features

The ADLX SDK contains the following components:

• ADLX interface library.
• Headers.
• Documentation and code samples.

Benefits

The ADLX SDK is a fast and easy way to develop differentiated applications that run on AMD systems.

• AI-assisted GPU management.
• Designed with quality and reliability at the core.
• Compatible with modern programming languages.
• Forward and backward compatible with AMD drivers.
• Global state synchronization.
• Access features and functionality in the areas of Display, 3D graphics, Performance Monitoring, GPU Tuning, and more!

Learn more by reading our blogs

ADLX 2.0: Extending graphics control to AI agents and agentic apps

AMD ADLX 2.0 adds AI extensions and MCP servers to help developers build intelligent apps that can monitor, manage, and optimize AMD graphics hardware.

RDNA 3: Introducing ADLX: The Next Evolution in Software Enablement

ADLX is a modern library and SDK designed to access features and functionality of AMD systems in the categories of Display, 3D graphics, and more.

Requirements

• Windows® 10 or later. 32 bit or 64 bit.

Footnote

For those looking for documentation on the legacy ADL SDK, you can find it here: ADL