AMD ROCm™ 6.4.4 on Radeon™ and Ryzen™ for Linux® Release Notes

Release Date:  September 24^th, 2025

This article provides information on the release of AMD ROCm 6.4.4 on Radeon and Ryzen for Linux®

New Product Support (Developer Preview)

• AMD Ryzen™ AI Max 300 Series
• AMD Ryzen™ AI 300 Series (Strix Point only) 

AMD ROCm 6.4.4 Highlights

• Developer preview support for PyTorch on AMD Ryzen™ processors

Installation Instructions

AMD Radeon™ graphics products users are recommended to install the Radeon™ Software for Linux® (RSL) 25.10.4 driver package along with AMD ROCm 6.4.4

• The Radeon™ Software for Linux® installer can be downloaded from the following link: Linux® Drivers for AMD Radeon™ and Radeon PRO™ Graphics
• For detailed instructions on installing AMD ROCm 6.4.4 with RSL, please refer to to ROCm on Radeon documentation

AMD Ryzen™ processor users are recommended to use inbox graphics drivers of Ubuntu 24.04.3 along with AMD ROCm 6.4.4

• For detailed instructions on how to install AMD ROCm 6.4.4, please refer to  ROCm on Ryzen documentation

AMD Product Family Compatibility

AMD ROCm 6.4.4 for Linux® is compatible with the following AMD products:

Compatible 64-bit Operating Systems

• Ubuntu 24.04.3 
• Ubuntu 22.04.5 (AMD Radeon™ series graphics products only) 
• RHEL 9.6 (AMD Radeon™ series graphics products only) 

Known Issues

• Increased memory consumption may be observed while running LLM FP8 inference workloads with PyTorch on Radeon™ RX 9060 XT graphics products.  
• Intermittent script failure may be observed while running Stable Diffusion FP16  inference workloads on some Radeon™ RX 7000 series graphics products.  
• Intermittent script failure may be observed while running Stable Diffusion inference workloads while suspending the system on AMD Ryzen™AI Max 300 series processors.  
• Intermittent script failure may be observed while running text-to-image inference workloads with PyTorch. 
• Lower than expected performance may be observed while running Llama 3 8B inference workloads with Llama.cpp. Users experiencing this issue are recommended to try an older version of Llama.cpp as a temporary workaround. 
• Intermittent script failure may be observed while running Stable Diffusion workloads with ONNX Runtime and MIGraphX. Users experiencing this issue are recommended to use MXR files instead of rebuilding the model as a temporary workaround. 
• Intermittent system or application crash may be observed while running Luxmark in conjunction with other compute workloads on Radeon™ RX 9000 series graphics products. Users experiencing this issue are recommended to shut down other applications and workloads while running Luxmark. 
• Intermittent script failure (out of memory) may be observed while running high-memory LLM workloads with multiple GPUs on Radeon RX 9060 serires graphics products. 
• Intermittent script failure may be observed while running BERT training workloads with JAX. 
• Intermittent script failure may be observed while running Stable Diffusion 2.1 FP16 workloads with JAX. 
• Intermittent script failure (out of memory) may be observed while running Llama2 FP16 workloads with ONNX Runtime and MIGraphX. 
• Intermittent system or application crash may be observed while running TensorFlow ResNet50 training workloads. 
• Increased memory consumption may be observed while running TensorFlow Resnet50 training workloads.

^{​​​© 2025 Advanced Micro Devices, Inc.  The information contained herein is for informational purposes only and is subject to change without notice. While every precaution has been taken in the preparation of this document, it may contain technical inaccuracies, omissions and typographical errors, and AMD is under no obligation to update or otherwise correct this information. ADVANCED MICRO DEVICES, INC. MAKES NO REPRESENTATIONS OR WARRANTIES WITH RESPECT TO THE ACCURACY OR COMPLETENESS OF THE CONTENTS OF THIS DOCUMENT AND ASSUMES NO LIABILITY OF ANY KIND, INCLUDING THE IMPLIED WARRANTIES OF NON-INFRINGEMENT, MERCHANTABILITY OR FITNESS FOR PARTICULAR PURPOSES, WITH RESPECT TO THE OPERATION OR USE OF AMD HARDWARE, SOFTWARE OR OTHER PRODUCTS DESCRIBED HEREIN. NO LICENSE, INCLUDING IMPLIED OR ARISING BY ESTOPPEL, TO ANY INTELLECTUAL PROPERTY RIGHTS IS GRANTED BY THIS DOCUMENT. This notice does not change the terms and limitations applicable to the purchase or use of AMD's products that may be set forth in a separate signed agreement between you and AMD.}

^{AMD, the AMD Arrow logo, AMD Radeon, AMD FirePro and combinations thereof are trademarks of Advanced Micro Devices, Inc.}

^{Linux® is the registered trademark of Linus Torvalds in the U.S. and other countries.}

^{Vulkan and the Vulkan logo are registered trademarks of the Khronos Group Inc.}

^{OpenCL and the OpenCL logo are trademarks of Apple Inc. used by permission by Khronos.}

^{OpenGL® and the oval logo are trademarks or registered trademarks of Hewlett Packard Enterprise in the United States and/or other countries worldwide}

^{PyTorch, the PyTorch logo and any related marks are trademarks of The Linux Foundation.}