Intel has been preparing to launch its next-generation Xe series GPUs, codenamed DG2 (Discrete graphics card 2).It is built by Intel DG1, Hoping to significantly improve performance so that it can compete with Best graphics card From AMD and Nvidia. The assumed lineup of Xe-HPG configurations ranges from 128 to 512 execution units (EU).Today, according to Leaked Geekbench entry, DG2-128EU may have a clock speed of 2,200 MHz.
All results, including the frequency of reports, should be carefully considered before we confirm them with official specifications. Nevertheless, compared with DG1, the clock speed is increased by about 40%, and the additional EU, higher memory bandwidth and architecture updates can make it reach a higher level of performance.
The real question will be how to stack the higher specification DG2 solutions. There are rumors that the DG2 series of Xe-HPG graphics cards may have up to five SKUs, and each SKU is configured with a different number of EUs and memory. Today’s leak is for the smallest, possibly budget-friendly solution.
Running on a test system consisting of an Intel Core i5-11400T Rocket Lake processor and 16GB single-channel (?) memory, the system scored 13,710 points in the OpenCL benchmark test. This is an entry-level performance segment, usually for low-power configurations. All of this is achieved with a GPU that purportedly runs at a clock speed of 2,200 MHz, which is very impressive for the new chip. Of course, Geekbench’s report may be wrong, so take this into consideration.
The test system is no different from our DG1 test bench, so we run it as a comparison point. It has a 65W i5-11400F. Compared with the i5-11400T used by DG2, it has a higher clock and higher TDP. We also ran the original single 8GB DIMM to further eliminate the advantages of the DG1 sample.This DG1 Geekbench 5 results Once again prove how bad Geekbench 5 is as a GPU benchmark, with a score of 19,176-40% higher than the assumed DG2 score of 13,710. A slower CPU may be part of the reason, but it is more likely that the early drivers and Geekbench 5 itself are the real culprits.
If we ignore the overall score and focus on the comparison of individual results, then DG2 does score from 25% higher to nearly 175% Than DG1 card.Basically not running Canny and Horizon detection tests will not help DG2, among which DG1 scores 23 times and 96 times higher, Obviously distort the overall image.
In any case, since Geekbench 5 obviously does not provide actual reliable performance estimates, we will have to wait for more leaks. Or postpone any ruling before Intel officially announces the Xe-HPG DG2 series of cards and independent testing. This will be the best way to determine real-world game performance. Assuming it has better drivers and higher clocks, its gaming performance may actually be usable.