AMD Vice President Hints at Possible ROCm Support for Windows Operating System

The expansion of AMD’s ROCm support from Linux to Windows could represent a significant advancement for both users and developers in the tech community.

AMD Confirms Potential for ROCm Support on Windows

For several years, there has been substantial anticipation surrounding the possibility of ROCm support on the Windows operating system. Users, particularly those involved in development, have expressed their eagerness for AMD to move forward with this initiative. Although AMD previously hinted at introducing ROCm functionality for Windows, progress has been slow, with the initial support offered for Windows 10 and 11 starting with version 5.5.1.

The latest version available is 6.2.4; however, it is essential to note that comprehensive support is limited to a select range of Radeon GPUs, primarily including specific models such as the AMD Instinct series and a few Radeon GPUs like the RX 7900 XT and XTX. Unfortunately, this means that not all Radeon GPU users can access ROCm on Windows, as support for the Radeon RX 9000 series remains lacking.

Yes

— Anush Elangovan (@AnushElangovan) March 7, 2025

In a recent statement, Anush Elangovan, AMD’s Vice President for AI Software, affirmed the company’s commitment to widening ROCm support on Windows. While details were sparse, his comments signal AMD’s readiness to extend the software stack compatibility to additional Radeon GPUs. Currently, users can deploy any RDNA 2 GPU on Linux without complications, contrasting the limited options available for Windows users.

ROCm versions Windows 10 and 11 — Image Source: AMD

While certain older models may function with ROCm on Windows, users may encounter significant compatibility and performance issues. Currently, the minimum eligible GPU is the RX 7900 GRE, which limits access to ROCm’s powerful features. Even with compatible GPUs, users might experience crashes, driver timeouts, and application freezes when running various tasks.

ROCm Support GPUs Linux — ROCm Supported Radeon GPUs on Linux

The challenges faced with ROCm on Windows are numerous and can be quite intricate. If AMD successfully addresses these issues and enhances support, users with older GPUs could unlock the ability to conduct deep-learning tasks. Moreover, expanding support to the latest RDNA 4 GPUs could unveil further potential, especially for developers utilizing Windows. Nonetheless, it appears that fully integrating ROCm support for Windows might still be some time away, given the lack of recent announcements from AMD.

CUDA isn’t really the moat people think it is, it is just an early ecosystem.tiny corp has a fully sovereign AMD stack, meaning we have rewritten the full stack from the hardware to PyTorch (with the exception of LLVM), and soon we’ll port it to the MI300X. You won’t even have to use tinygrad proper to use it, tinygrad has a torch frontend now.

Either NVIDIA is super overvalued or AMD is undervalued. If the petaflop gets commoditized (tiny corp’s mission), the current situation doesn’t make any sense. The hardware is similar, AMD even got the double throughput Tensor Cores on RDNA4 (NVIDIA artificially halves this on their cards, soon market pressure will force them not to).

I’m betting on AMD being undervalued, and that the demand for AI has barely started. With good software, the MI300X should outperform the H100.

In a related development, tinygrad is set to receive two MI300X systems from AMD, marking a notable progression for the AI development sector.

AMD 💕 @__tinygrad__

we are looking forward to working closely with @__tinygrad__ to help commoditize the petaflop https://t.co/LEjsUaPWHV

— Anush Elangovan (@AnushElangovan) March 7, 2025

The developer community believes that if AMD can successfully enhance its software offerings, it could create a more level playing field against NVIDIA’s market dominance, which currently values NVIDIA significantly higher than AMD.

Source & Images