Intel is planning an assault on the premium end of the GPU market, if recent rumours are to be believed. Latest murmurings point to a much larger GPU for next-gen Battlemage architecture, boasting double the shaders compared to Intel’s top dog of today, A770.
According to YouTube channel RedGamingTech, Intel Arc’s second GPU family – codename Battlemage – is to feature a monolithic die hosting 64 Xe cores clocked at 3GHz or more, 48MB of L2 cache, and a 256-bit bus connecting GDDR6 (maybe X) memory; all inside a similar die size to Nvidia’s AD103 GPU, which is a large chip sitting at 379mm² in size and comprising 45.9bn transistors. You’ll know this one as GeForce RTX 4080 on the desktop.
Finally, there are two notes about improved ray tracing performance and an overhauled Xe-core design, manufactured using TSMC’s 4nm node.
But what do these numbers mean for gamers? Well, again, according to RedGamingTech’s sources, Intel is targeting RTX 4070 Ti to RTX 4080 performance range, which makes sense given the more muscular design. Will Intel manage to hit this tier is unknown, especially as there’s more to a GPU than mere silicon alone.
With that said, looking at how Intel’s drivers are becoming better with each update, there is a chance for them to become as good as their competitors by the time Battlemage reaches store shelves.
While some may criticise Intel for not offering a truly high-end product like Nvidia’s RTX 4090 from the get go, keen pricing puts today’s Arc in a good spot. Not everyone needs a 450W graphics card that costs more than some second-hand cars.
At the end of the day, the better competition becomes, higher are the generational leaps, and lower the pricing gets. When many players are in the game, even if two play ball with each other, there will be a third ready to shake things up, benefiting the end user no matter the chosen brand. Bring on the ‘mage!