A few days ago, we told you about the "unofficial" confirmation of the existence of a renewed GeForce RTX 3060 Ti equipped with GDDR6X memories and no longer the GDDR6 ones like the original edition. Now, NVIDIA has decided to officially announce the card, which will also be used by the company to get rid of the stock of Ampere GPUs, so as to focus mainly on new products based on the Ada Lovelace architecture (RTX 40 series). Not only that, there will also be a GeForce RTX 3060 model equipped with 8GB of GDDR6 memory. The California-based company said:
To give gamers and creators more choice, we are introducing a couple of additional options from our network of graphics card partners around the world, starting in November. The first is a faster GeForce RTX 3060 Ti with GDDR6X memory, in addition to the original version with GDDR6 memory. The second is a GeForce RTX 3060 with 8GB of GDDR6 memory. Both new options are supported starting with today's Game Ready driver. For more details on each option, visit the GeForce RTX 3060 family product page on GeForce.com.
Photo Credit: NVIDIA
The RTX 3060 8GB maintains the same number of CUDA cores and boost clocks as the 12GB edition, i.e. 3584 and 1,770MHz respectively, but now has 8GB on a 128-bit bus compared to 12GB on a 192-bit bus of the original. This results in a 33% reduction in bandwidth, plus the obvious decrease in capacity.
From a performance standpoint, the new RTX 3060 Ti GDDRX6 should come very close to an RTX 3070, while the RTX 3060 8GB will obviously have lower performance than the 12GB variant, but also a lower price. For the occasion, NVIDIA has also released a new release of its graphics drivers, which you can retrieve directly from the official website.