Radeon Limitations and recommended settings

7.2.1 release known issues#

Linux#

Known issues#

  • Application or script failures may be observed when running ONNX Runtime workflows with the MIGraphX Execution Provider on some AMD graphics products, such as the Radeon™ AI PRO R9700

  • Failures may be observed when running Janus Pro inference with transformers==5.1.0 on some Radeon™ AI PRO series graphics products, such as the Radeon™ AI PRO R9700. Users experiencing this issue are recommended to use transformers==4.57.5 as a temporary workaround.

  • GPU memory access fault errors may be observed when running Unsloth QLoRA fine‑tuning for Llama 3 8B on RHEL with Radeon™ AI PRO R9700 using ROCm‑based PyTorch.

  • Abnormal colors may be observed when running the ComfyUI Wan2.2 TI2V 5B model on some Radeon graphics products, such as the Radeon™ AI PRO R9700.

Limitations#

  • AO Triton with PyTorch 2.9 is disabled by default for AMD Radeon RX 7000 series graphics products, and must be enabled manually.

Multi-GPU configuration#

AMD has identified common errors when running ROCm™ on Radeon™ multi-GPU configuration at this time, along with the applicable recommendations.

See mGPU known issues and limitations for a complete list of mGPU known issues and limitations.

Windows#

Note
The following Windows known issues and limitations are applicable to the 7.2 release. Only Pytorch is currently available on Windows - the rest of the ROCm stack is only supported on Linux.
AMD is aware and actively working on resolving these issues for future releases.

Note
If you encounter errors related to missing .dll libraries, install Visual C++ 2015-2022 Redistributables.

Known issues#

  • Disable the following Windows security features as they can interfere with ROCm functionality:

    • Turn off WDAG (Windows Defender Application Guard)

      • Control Panel > Programs > Programs and Features > Turn Windows features on or off > Clear “Microsoft Defender Application Guard”

    • Turn off SAC (Smart App Control)

      • Settings > Privacy & security > Windows Security > App & browser control > Smart App Control settings > Off

Limitations#

  • No ML training support.

  • Only Python 3.12 is supported.

  • For ComfyUI, adding the --lowvram and --disable-pinned-memory parameters may help with lower-memory configs.

  • On Windows, only Pytorch is supported, not the entire ROCm stack.

  • On Windows, the latest version of transformers should be installed, via pip install. Some older versions of transformers (<4.55.5) might not be supported.

  • On Windows, only LLM batch sizes of 1 are officially supported.

Contents