With the absolute lack of stock by AMD and NVIDIA graphics cards due to the mining boom, Intel is not interested in taking out its Intel Xe-HPG GPUs even if it wants to earn the trust of the most enthusiastic and stop being the duckling. ugly of gaming GPUs.
Intel does not want to be below AMD and NVIDIA
The last time Intel released a dedicated graphics card for gaming was with its failed i740 and i750 in the late 1990s, so it’s been more than twenty years since then. Not for that reason Intel completely abandoned the market for graphics architectures, but it has always focused on low-end integrated graphics, which have never been good enough, knowing this problem they decided a few years ago to solve it by improving their graphics architectures so that they are competitive against NVIDIA and AMD.
The result has been a graphic architecture that is going to be launched in two different generations, where the first of them has gone to the lowest range on the market. Which are the integrated graphics in its latest Tiger Lake and Rocket Lake-S CPUs, as well as the Intel Xe Max, but none of these GPUs have a powerful enough architecture and are designed to operate at low power consumption. Hence its name Intel Xe-LP.
The second generation instead is referenced in three different models: Intel Xe-HP, Intel Xe-HPC and Intel-Xe-HPG. In the case of Intel Xe-HP and Xe-HPC they are GPUs designed for professional markets outside the home, and therefore will compete against NVIDIA Tesla and AMD CDNA. On the other hand, the Intel Xe-HPG are its first gaming GPU in twenty years.
First images of the Intel Xe-HPG
Intel has already finished the prototype or engineering sample of its graphics card for gaming, the Intel Xe-HPG, which we know thanks to the fact that its first images have been leaked. So we can conclude that its launch is close in time. The engineering samples are complete prototypes of a fully functional graphics card and are built after the graphics chip they will house has been completed.
Information about the Xe-HPG has been leaked from the YouTube channel “Moore’s Law is Dead” and it shows a graphics card that uses a two-fan system. At the same time they have given information about their specifications, which are the following:
- Configuration of 512 EUs, this equates to 64 Compute Units in AMD RDNA 2 in terms of the number of compute units available.
- It will be manufactured under TSMC’s 6 nm node, it will not be Intel who makes its gaming GPU since they do not want to saturate their wafers even more, especially when they are going to make the leap to their 10 nm SuperFin process in both notebook CPUs and desktop like Alder Lake.
- 256-bit GDDR6 bus.
- 2.2 GHz speed and consumption less than 300 W, it is spoken of 275 W.
Due to its specifications, it would be in the mid-range range, AMD Radeon RX 6800 and RTX 3070, although who has leaked the information speaks of performance around an RTX 3070 Ti. Nor can we forget that Intel could launch a full range with configurations with fewer EUs and lower GDDR6 bandwidths.