NVIDIA GeForce RTX 3060 Ti promises to be the most affordable new graphics card in 2020. And it’s not about the price at all, although it is really lower than that of other new products from NVIDIA and AMD, but about the availability on the shelves. According to retailers around the world, they received significantly more GeForce RTX 3060 Ti units from suppliers than they had the rest of the GeForce RTX 30-series graphics cards at the time of their launch.
Hardware Canucks surveyed 12 different retailers in the United States, Canada, Europe and Australia. Most of the sellers said that on the eve of the start of sales, they formed significant stocks of GeForce RTX 3060 Ti, comparable to the total number of GeForce RTX 3070, RTX 3080 and RTX 3090 video cards that they had at the time of launch. Some noted, however, that the RTX 3060 Ti’s shipments were comparable to the RTX 3070.
As a result, it turns out that the start of sales of the GeForce RTX 3060 Ti turned out to be better prepared than in the case of any other video card released over the past few months. A particularly striking contrast is seen when compared to the recent disastrous launch of AMD’s Radeon RX 6800 and RX 6800 XT. Recall that there were so few reference models of cards based on AMD chips that they did not reach many stores at all, and the start of sales of non-reference versions took place only formally, but not actually.
NVIDIA has priced the GeForce RTX 3060 Ti at $ 399. The novelty of the Ampere family surpasses the previous gaming flagship GeForce RTX 2080 Super, which cost $ 699, in performance in games and other tasks. This gives the GeForce RTX 3060 Ti every chance of success, especially given the lack of other offers. True, NVIDIA partners have yet to confirm their ability to maintain retail inventories.
GeForce RTX 3060 Ti is built on the Ampere GA104-200 GPU, which has 4864 CUDA cores, 152 tensor cores, 38 RT cores, 152 texture units and 80 rasterization units. They provide peak performance up to 16 teraflops in shader operations, up to 32 teraflops when using RT cores and up to 130 teraflops in tensor calculations.
If you notice an error, select it with the mouse and press CTRL + ENTER.