Intel is making a bold statement about its commitment to the discrete GPU market with the unveiling of its new Xe2 ‘Battlemage’ architecture. This architecture, set to make its debut in Lunar Lake-based laptops later this year, promises a significant 50% performance upgrade over its previous generation, Xe ‘Alchemist’.
During a recent episode of Intel’s ‘Talking Tech’ series, Tom Peterson, a key figure in the company’s graphics development, outlined the key features of Battlemage. He emphasized the scalability of the Xe2 architecture, which will power both integrated graphics in the Lunar Lake mobile processors and the next-generation Arc series of discrete graphics cards.
Intel has been focusing on improving performance per watt and per area with Xe2, a strategic move to optimize both its integrated graphics for power-constrained devices and its discrete graphics for performance-focused desktops. This emphasis on efficiency is evident in the memory configuration of the Battlemage GPU.
The initial Battlemage GPU configuration boasts a 192-bit memory bus with 12GB of GDDR6 memory operating at 19Gbps, delivering a peak bandwidth of 456GB/sec. While this represents a significant improvement over the 17.5Gbps GDDR6 on Arc Alchemist, which featured a 256-bit memory bus, Intel plans to offer higher-end Battlemage GPUs with a 256-bit memory bus and 16GB of GDDR6, achieving a bandwidth of up to 608GB/sec.
The launch of Battlemage in Lunar Lake laptops marks the first step in Intel’s ambitious strategy to re-enter the discrete GPU market. Following its debut in mobile devices later this year, Battlemage will make its way to desktops in 2025, facing stiff competition from established players like NVIDIA and AMD.
With its focus on performance, efficiency, and scalability, Intel’s Battlemage architecture represents a significant step forward in its graphics ambitions. The coming months will be crucial for Intel, as it seeks to establish its place in the highly competitive discrete GPU market.
