AMD ROCm is about to land on the Windows operating system
Recently, AMD shared two significant announcements within the ROCm community: the imminent arrival of the ROCm SDK on Windows operating systems and AMD’s intention to extend support for consumer-grade Radeon graphics cards.
In 2016, AMD introduced ROCm (Radeon Open Compute Ecosystem), an open software platform conceived as an open-source alternative to NVIDIA’s CUDA platform. Built for high flexibility and performance, ROCm targets accelerated computing without language constraints, enabling participants in the machine learning and high-performance computing communities to expedite code development through various open-source computing languages, compilers, libraries, and redesigned tools. The platform is suitable for large-scale computing, supports multi-GPU calculations, and is compatible with CDNA and RDNA architecture GPUs.
Initially, the ROCm support list was limited to specific Instinct and Radeon Pro models, including Instinct MI250X, MI250, MI210, MI100, and MI50 as fully supported. Radeon Pro W6800 and Radeon Pro V620 were also listed. AMD has now expanded the list to encompass Radeon RX 6900 XT, Radeon RX 6600, and Radeon R9 Fury. However, only Radeon R9 Fury boasts complete software-level support on the ROCm platform, while the other two RDNA 2 architecture products offer partial support.
Originally, AMD designed ROCm for use with Linux. However, alternative methods, such as Docker or the Windows Subsystem for Linux (WSL) virtualization techniques, enabled its operation on Windows-based systems. Although performance is somewhat affected compared to native implementation, the introduction of ROCm support for Windows operating systems fulfills a long-standing demand for many users.
By broadening the ROCm ecosystem, AMD continues to progress in the right direction, though much work remains to be done in refining the platform.