### AMD ROCm 6.4 Still Lacks Support for RDNA 4 Architecture Over a month since the launch of the Radeon RX 9000 series of graphics cards, AMD's Radeon Open Compute (ROCm) software stack still does not officially support the RDNA 4 architecture. This is a significant setback for developers and researchers who rely on ROCm for high-performance computing (HPC) and artificial intelligence (AI) applications. ROCm, an open-source software stack, is designed to optimize GPU performance across various platforms, making it a popular choice among users. However, the lack of support for the latest generation of AMD GPUs has sparked widespread disappointment and discussion within the community. The release of ROCm 6.4 brought several improvements, including optimizations for existing architectures and enhanced support for various AI frameworks like TensorFlow and PyTorch. Despite these advancements, the absence of RDNA 4 support is a glaring omission. AMD has acknowledged the issue and stated that their development team is working diligently to address it, but no specific timeline has been provided. Users are increasingly turning to third-party drivers and early versions of ROCm to gain support for their new cards, reflecting the urgency of the situation. For AMD, introducing RDNA 4 support into ROCm is crucial for maintaining its competitive edge in the HPC and AI markets. Early support would demonstrate AMD's commitment to developers and researchers, potentially expanding its reach in academic and commercial sectors. While the company faces challenges, it remains committed to resolving these issues in future updates. ### Key Innovations in ROCm 6.4 The latest version of ROCm, 6.4, marks significant progress in several areas, further solidifying AMD's position in the AI and HPC domains. Firstly, ROCm 6.4 has achieved major improvements in model inference, with notable performance boosts in multi-node clusters. This enhancement is particularly beneficial for large-scale deployments, providing faster response times and maintaining high performance even under complex tasks. Secondly, AMD has introduced optimized containerization tools, making the transition from development to production more seamless for AI applications. This "plug-and-play" approach helps developers deploy and manage applications more efficiently, reducing configuration and compatibility issues. Thirdly, ROCm 6.4 includes modular deployment tools that offer greater flexibility in managing large GPU clusters. Users can dynamically adjust cluster configurations based on their specific needs and resource availability, leading to higher resource utilization and better performance optimization. Additionally, ROCm 6.4 has enhanced support for leading AI frameworks such as TensorFlow and PyTorch, allowing for seamless switching between frameworks. This increases the flexibility and portability of AI development. The platform also features more detailed performance analysis tools, aiding developers in optimizing model training and inference processes. User experience has been significantly improved with ROCm 6.4. The installation and configuration processes have been simplified through automated tools and enhanced documentation, making it easier for new users to get started. Furthermore, AMD has strengthened community support, offering more resources and tools to help users troubleshoot and share experiences. ### AMD's Move into the Budget Market with RX 9060 XT In a bid to expand its presence in the budget market, AMD has announced the upcoming release of the RX 9060 XT graphics card. This new offering will be available in two versions: 8GB and 16GB. The RX 9060 XT, which uses the Navi 44 core, aims to compete with Nvidia's RTX 5060 Ti, catering to the needs of mid-to-low-end users. The RX 9060 XT is expected to have a core frequency of 3.2 GHz, a high figure in the current market, which will significantly enhance both gaming and computational performance. The 16GB version is particularly well-suited for handling demanding graphical tasks and large-scale games, while the 8GB version targets budget-conscious users. AMD is known for delivering high-performance products at competitive prices, and the RX 9060 XT is no exception. The Navi 44 core not only represents a more advanced fabrication process but also offers better energy efficiency and thermal performance. The release date for the RX 9060 XT is expected in the coming months, although the exact timing remains unspecified. This announcement has garnered significant attention from consumers, especially those interested in the value proposition of AMD's offerings compared to Nvidia's. ### Industry Reactions and AMD's Market Position Industry insiders have praised the release of ROCm 6.4, highlighting its potential to close the gap between AMD and NVIDIA in the AI market, particularly in HPC and large-scale model inference. The improvements in inference performance, containerization, and modular deployment tools are seen as significant steps forward. AMD's continued investment in its ROCm platform demonstrates its dedication to the AI and HPC communities, reinforcing its position as a leader in these fields. In terms of the RX 9060 XT, the card's high core frequency and advanced Navi 44 core make it an attractive option for both gaming and professional use. AMD's focus on the budget market with a high-performance product aligns with its strategy to offer more accessible solutions and challenge Nvidia's dominance. The RX 9060 XT's release is expected to expand AMD's customer base and provide more options for consumers. ### Company Profile AMD, a global leader in semiconductor technology, is renowned for its innovations in high-performance computing, graphics, and AI. The company's dedication to open-source software and advanced hardware solutions has made it a formidable player in the tech industry. With initiatives like ROCm and the introduction of budget-friendly GPUs like the RX 9060 XT, AMD is positioning itself to compete effectively in both the high-end and mid-to-low-end markets, driving innovation and providing users with versatile, high-performance options.