Intel has unveiled that it has started to sample its Iris Xe HPG DG2 with board partners. Intel made the announcement at ISC 2021 that the Xe HPG platform, previously known as DG2, is now sampling to partners.
From the looks of it, Intel’s Xe HPG platform is getting into the final leg of the journey of sampling and production. Intel also restructured the company to put graphics front and center and created a new business unit, the Accelerated Computer Systems and Graphics (ACSG) Group, led by the Intel GPU Chief Raja Koduri, last week. The ACSG is like the ‘Radeon’ sub-brand of Intel.
A few weeks ago, it was leaked that the DG2-512 GPU die is not as big as the flagship GPU offerings from NVIDIA and AMD. But if we are to compare the high-end GPU category (Navi 22 / GA104), the DG2-512 has a slightly bigger die than the other two.
The DG2-512 GPU has a BGA-2660 package which measures 37.5mm x 43mm. In comparison, NVIDIA’s Ampere GA104 measures 392mm2 which means that the DG2 chip is comparable in size while the Navi 22 GPU measures 336mm2 or around 60mm2 less.
NVIDIA’s GPUs pack tensor cores and much bigger RT/FP32 cores in their chips while AMD RDNA 2 chips pack a single ray accelerator unit per CU and Infinity Cache. Intel is expected to have hardware-accelerated ray-tracing capabilities onboard it’s Xe-HPG GPUs as well.
A benchmark for the Intel Xe-HPG DG2 GPU also leaked out a while ago which gave a detailed idea of what specifications the Xe-HPG DG2 will have. The Intel Xe-HPG DG2 GPU spotted in the benchmark has 512 EU (Execution Units) which rounds up to 4096 cores. The Xe-HPG DG2 GPU was running at a clock speed of 1800 MHz (max) in the benchmark. We do not know what the final clock speeds will be for the GPU. The leaked benchmarks also detail that Intel Xe HPG DG2 will almost match Nvidia’s RTX 3080 in performance.
We will have to wait and see what Intel has in store with its graphic offerings.
