The RTX 4070 has not reached the market on the right foot despite the fact that NVIDIA has made its efforts to offer it at a lower price than the one that was leaked two months ago. The 599 dollars are not enough to satisfy the public, so the next option is already in the chamber, which we hope has a more restrained price and suitable for the times. We are talking about the RTX 4060 Ti, which has revealed its frequency base and with overclock, as well as its performance in FP32 and TFLOPS.
The rumors are taking shape and somehow they are overlapping, confirming each other. Behind the scenes, there was talk of higher frequencies than expected for this model, not by much, yes, but it would be one more step for the greens to get dangerously close to the 3 GHz figure completely as standard.
NVIDIA RTX 4060 Ti, base frequencies and overclock
It will not disappoint in this section, and possibly not in the final performance either, which we will see later. the leaker T4C Fantasy has leaked two key data such as the frequency that the RTX 4060 Ti will have in its Founders Edition or reference version (it is not clear if NVIDIA will launch an FE model as such) and with overclock.
First of all, the figure that we already saw months ago is confirmed again: 2.535 MHz for the so-called Boost Default, that they should have all the models that follow the guidelines of the green ones. There are no surprises here, however, there are on the factory overclocked models.
Specifically, we talk about the Colorful iGame RTX 4060 Ti Ultra W OCa model that is assumed to be the top of the range within the manufacturer, which will arrive with some extra MHz, specifically 45MHz
to reach the 2,580MHz, a figure that is not surprising at all for the model that it really is. Be careful because this is an important performance “layer” if we take into account the previous leaks of the same leaker.
VRAM and performance
There are not too many surprises either, since the 8GB that this model will have, remaining, in theory, again behind AMD in this section at the same generation. This would indicate that it would be used again a larger L2 cache to use with it a smaller bus, which in this case, is 128 bitachieving better efficiency per watt than in the RTX 30.
Regarding performance, a figure close to 22 TFLOPS in FP32 for the model with frequencies Founders Editionand ones 24 TFLOPS for models with overclock serial. In comparative terms, the RTX 4060 Ti will be in its versions with reference frequencies at the height of an RTX 3070 Tiwhile overclocked versions should approach or match an RX 6900 XTXstaying a little below an RTX 3080 FE. Regarding the difference with its predecessor (RTX 3060 Ti) it should go between 32% to 37% in the fastest custom models.
To finish, TBP and price. In the first of them, it is estimated that he will be among the 160 watts for models referenceand the 180 watts for models with overclock, where it seems that the PL will be similarly limited from overlapping the RTX 4070 in conditions of high voltage and increased frequencies. Regarding the most controversial section of this series, the price, a cost of $399that at the change and already with taxes in our country should be among the 440 euros and 450 euros.