Not too long ago, if they had told us that one of NVIDIA’s releases for the next few months would be a new version of a previous generation RTX 20, we wouldn’t have believed it because of the existence of the RTX 3050. But in the end it does and they have disclosed the (corrected) specifications of the 12 GB RTX 2060, so it will be a reality. What can we expect from them?
Where does the existence of 12 GB RTX 2060? Keep in mind that this is a graphics card whose main GPU is manufactured under the TSMC “12 nm” process and therefore does not have to share wafers with the RTX 30 range manufactured by Samsung. in 8 nm. Add to this that most of the chips for the RTX 3050 (Ti) are being installed for gaming laptops and Intel will be releasing an entry level Intel ARC and AMD has its Radeon RX 6500 XT, then it’s very clear what it is about the reason for the existence of this graphics card.
The 12GB RTX 2060 is a consequence of the lack of production chips for a desktop RTX 3050 (Ti), but mostly because the RTX 2060 achieves higher benchmarks than the desktop version of the RTX 3050. Ti , which caused NVIDIA to make a 180º turn in its initial strategy and decide to relaunch the RTX 2060, up to 12 GB of VRAM.
Specifications 12 GB RTX 2060
Looking at the specs of the 12GB RTX 2060, it can be said that it is halfway between the original model and its later SUPER model. From the outset the 12 GB indicates a 192-bit bus using 2 GB memory chips, a step back from the SUPER version and its 256-bit bus. Bandwidth ? 336 Gbps, so the speed is still 14 Gbps.
As for the GPU itself, it’s exactly the same as the RTX 2060 SUPER, with its 2176 ALU in FP32 in 34 SM of Turing architecture. Which gives you a maximum power of 8.1 TFLOPS and FP32
So at the end of the day we are faced with an RTX 2060 SUPER with the VRAM cut off and therefore we should expect much lower performance as the bandwidth has not increased so it should not be much better than the current RTX 2060, which due to the insane madness of mining retails for prices close to 600 dollars, despite its low profitability.
A little higher consumption
Despite the cut in terms of VRAM, it seems that this variant of the RTX 2060 will consume 185W, something surprising if we take into account the 175 W of the SUPER model. The only thing we can think of is that NVIDIA lowered the efficiency in order to fend off miners. Is this the strategy to stop the mining boom?