AMD Radeon RX 7900 XT supports ROCm 5.7 for PyTorch ML and AI acceleration

Last month, AMD released the ROCm 5.7.1 driver update for Ubuntu Linux, extending ROCm and PyTorch support to the Radeon Pro W7900 and Radeon RX 7900 XTX graphics cards. Before the commencement of the AMD Advancing AI event in early December, AMD announced the availability of ROCm and PyTorch support for the RX 7900 XT graphics card, which also utilizes the RDNA 3 architecture, thereby encompassing all the top-tier products within the RDNA 3 range.

AMD’s Software Product Manager, Erik Hultgren, expressed enthusiasm about the newest additions to their product portfolio. These high-end GPUs, in synergy with ROCm, simplify the utilization of artificial intelligence from both a software and hardware perspective, thus enabling developers to select the solutions that best align with their requirements.

Like the other two graphics cards, the RX 7900 XT boasts an impressive 20GB of memory and houses a total of 168 AI acceleration units on the Navi 31 chip, offering rapid machine-learning training and inference capabilities. The GPU also supports PyTorch-based machine learning models and algorithms, which have been optimized specifically for AMD and are compatible with the Ubuntu Linux system running version 22.04.3.

It is hoped that this is just the beginning, as AMD possesses a plethora of other RDNA 3 architecture GPUs that consumers are eager to see supported by ROCm. In comparison, their competitors, NVIDIA and Intel, have been providing superior general computing support on consumer-grade graphics cards, such as CUDA and the open-source Intel Compute Runtime. In contrast, AMD’s current support for ROCm is considerably limited.