When it comes to Intel’s naming convention of its Xe GPU lineup, we can honestly think of a few names that are less confusing. That aside, it is being reported that an allegedly new GPU for its Xe-HPG lineup has cropped up and made an appearance on Geekbench.
To be precise, the Xe-HPG GPU in question appears to have a maximum clockspeed of 2.2GHz, 128 Compute Units or, in this case, Execution Units (EUs), and an odd 4.67GB of memory. In addition, the alleged Xe-HPG GPU was seemed to have been paired with an Intel Core i5-11400T and a B560M-PLUS motherboard from ASUS.
Given the number of EUs, it should be assumed that the GPU listed here could be one of Intel’s lower-tier, entry-level DG2 GPUs, given the number of EUs being housed. As previously reported, the highest tier variant of the Xe-HPG variant will reportedly house up to 512 EUs; one particular SKU with 448 EUs reportedly gives NVIDIA’s GeForce RTX 3070 a run for its money, as well as AMD’s Radeon RX 6700XT.
In the case of this 128EU-powered Xe-HPG DG2 GPU, its OpenCL score of 13710 points tells us that the GPU may quite possibly be on par with a GeForce GTX 660 Ti or Radeon RX 550.
Sadly, as things usually go with Intel’s Xe GPUs, there still isn’t any concrete performance metrics for any of the GPUs. It also doesn’t help that Intel’s lips on the matter have been air-tight. On that note, looks like we may still have to wait until CES 2022 to find out more.
(Source: Geekbench, Videocardz)
