The return to graphics cards for gaming by Intel is one of the events in terms of hardware that arouses the most enthusiasm, not only for having been absent for 20 years after the fiasco of the i740 at the end of the 90s, but for the fact that this will mean the end of the duopoly between NVIDIA and AMD in all ranges.
This includes the entry-level range of GPUs for gaming, which has been somewhat neglected by NVIDIA and AMD for a while and is going to be in that market segment where Intel will seek to gain market share. Which has made his rivals have made defensive movements. Well, the GPU that we used to know as DG2-128 now has a name: ARC Alchemist A380.
Intel ARC A380 Specifications
The most modest of Intel's first-generation graphics architecture, Alchemist, will be called the A380 and we have to say that it caught us by surprise, as we expected a name like A350 to match AMD and NVIDIA alternatives such as RX 6500 XT and the still to be confirmed RTX 3050 (Ti) for desktop although it did launch on laptops.
Its technical specifications? We are talking about 1024 ALUs in FP32 distributed in 8 Xe Cores and a clock speed of 2.45 GHz, which has not yet been revealed to us if it will be in Boost mode or the usual clock speed while we play (probably the second). In any case, it seems that its performance is that of a GTX 1650 SUPER and is accompanied by a 96-bit bus that allows it to place 6 GB of GDDR6 VRAM on the card, of which we do not know its final speed, but everything points to than the greater margin that a desktop GPU gives compared to a laptop will be used to place the video memory at 16 Gbps.
As it is a graphics card in the entry range, it will be enough with a PCI Express port without an additional connector for power and, therefore, its consumption will be 75W or less.
What can we expect from Intel GPUs?
The fact that Intel is betting on the input range of gaming graphics is good news in principle, despite the fact that the A380 is put wide with an NVIDIA Turing capable of the GTX 16, which have two Years old when NVIDIA and AMD have more efficient architectures make us wonder if Pat Gelsinger's will measure up or if they have an ace up their sleeve that they have not yet revealed.
The GTX 1650 SUPER has a 1280 ALU configuration, a 128-bit GDDR6 bus, a 100W TDP and a clock speed between 1,530 MHz and 1,725 MHz. In general, Intel's A380 is much more efficient than the old-fashioned one. NVIDIA design, but we expected higher performance and since the higher ranges share architecture makes us somewhat skeptical of the performance of the higher ranges.
With the shortage of graphics cards to play the market is with open arms, willing to receive any GPU as long as they can buy it to mount them on their PC. Although we believe that a good part of these A380s will go to several pre-built PCs first.