ROCm 6.1.3 supports multiple GPUs and the Windows subsystem for Linux

AMD continues to expand the opportunity for use Artificial Intelligence outside the data centers enabled for this type of tasks, thanks to its ROCm software. This software Open Source allows to use home use graphics cards for AI loads and thus offer small developers a solution alternative. The AMD ROCm version 6.1.3 expands these possibilities even further by becoming compatible with Windows subsystem for Linux among other news.

This new version of AMD ROCm 6.1.3 comes with some interesting new features, now it is compatible with multi-GPU configurationswhich will allow the creation of solutions for multiple services thanks to scalability. Also offered in beta level support for the Windows subsystem on Linuxspecifically with WSL 2Which will allow execute Linux AI tools on a Windows system. It is also done compatible with TensorFlowadding another option to the already existing PyTorch and ONNX.

With this new version they have taken the opportunity to include the new card AMD Radeon Pro W7900 Dual Slot and what it offers 48GB VRAMan ideal solution for adding a multi-GPU configuration to AMD workstations for large language models LLM such as LLaMA-2 or LLaMA-3, which require at least 35 GB of GPU memory.

End of Article. Tell us something in the Comments!