Intel Reveals Simplified Next-Gen Arc Battlemage GPU Strategy In Interview
Enthusiasts who've been around for a little bit will recall that before Arc existed as a brand, all of Intel's talk around its graphics hardware was about Xe, the core architecture for the Arc brand. Xe was actually split into no less than four separate architectures: Xe-LP, Xe-HP, Xe-HPG, and Xe-HPC, all targeted at different segments of the market.
Intel's original vision for Xe included many different architectures.
Intel's not taking that tack with Xe2, as it turns out. Now that the Accelerated Compute Group (AXG) has been split into two sections specifically serving Intel's Client Computing Group and Data Center Group, it looks like the next generation of Arc graphics will simply be split into two segments: Xe2-LPG for integrated graphics and low-power parts, and Xe2-HPG for powerful discrete GPUs and pro visualization.
This news comes along from an interview with Intel gave over at German site HardwareLuxx, where Intel's Tom "TAP" Petersen outlined in vague terms the future of Intel Xe after Alchemist. It looks like the Data Center Group will focus on revising Xe-HPC, while the slice of AXG that joined the Client Computing Group will be focusing on the two flavors of Xe2.
Intel's original plan had Battlemage, the second-generation Arc graphics architecture, launching in 2022. Continual delays due to process and execution issues led to Arc making a very late debut on the desktop in Q3 of last year. Intel hasn't said when Battlemage will launch, but the latest leaked roadmaps from the company seem to indicate that it won't show up until the beginning of next year.
