Intels Next "Battlemage" GPU from Arc Series to Feature Double Xe-Core Count and TSMC 4nm-EUV Node

According to a recent report by RedGamingTech, Intel's highly anticipated next-generation "Battlemage" GPU from the Arc series is expected to bring a significant improvement over the current ACM-G10 GPU by doubling its shader counts. 

The rumors suggest that the Xe cores will be doubled, resulting in a staggering 8,192 FP32 shader units, which will significantly improve the performance of the graphics cards. Intel is taking the gaming world by storm with the new Arc graphics cards, and we can expect great things from this new release.

The upcoming high-end "BGM-G10" GPU will compete in the high-performance segment and is speculated to feature a TSMC 4nm-class EUV node, similar to NVIDIA's GeForce "Ada" GPUs. Moreover, the die-size of the "BGM-G10" GPU is expected to be similar to that of the "AD103" silicon, which powers NVIDIA's GeForce RTX 4080.

As per the latest rumors, the flagship Battlemage GPU, likely to be named BGM-G10, will come with up to 64 Xe-Cores, twice the count of its predecessor. Additionally, it is expected to have a 256-bit memory bus with 48MB of L2 cache, which would allow graphics cards based on this processor to feature 8, 16, or 32GB of memory, depending on whether Intel uses a 256-bit or cut 192-bit memory bus. Although the L2 cache is smaller than the RTX 4080's AD103 (64MB), the die size is said to be similar.

While the clock target for the Battlemage GPU is around 3GHz, it is important to note that Intel's GPU clock is not the same as NVIDIA's Boost Clock, and the RTX 40 series often hit 2.9GHz with under light workloads and overclocking. Although it is too early to confirm, speculation suggests that the BGM-G10 could perform at least at the RTX 4070 Ti level, making it a significant upgrade over the Arc A770, which targets NVIDIA's RTX 3060 series.