For their testing, the Polish tech site ITHardware got their hands on the first RTX 3060 Ti graphics card with GDDR6X memory.
In October of last year, NVIDIA released an improved version of its RTX 3060 Ti model. The Polish site currently has access to the KFA2/GALAX RTX 3060 Ti G6X OC variant and compared it to RTX 3060 Ti and RTX 3070 cards.
The new RTX 3060 Ti is installed with the same GA104 GPU and 4864 CUDA cores as its predecessor. The only significant change is the RAM increase from 14Gbps GDDR6 to 19Gbps GDDR6X. This indicates that the maximum potential bandwidth has grown from 448 GB/s to 608 GB/s, albeit at the expense of higher power consumption (200W to 225W).
The performance boost is negligible, ranging from 1% to 2%, but there is an important point that must be noted first. The ASUS RTX 3060 Ti TUF model utilized for comparison (based on GDDR6 RAM) is a factory-overclocked variant with a boost upgraded from 1665 to 1755 MHz, which is an increase of 5.4%. Thus, this is not an ideal comparison since such testing should be conducted across models with comparable specifications.
Nonetheless, this indicates that the new card is quicker than its factory-overclocked predecessor. Although the growth is not double digits, it is also not 1-2%, as the figures may initially imply. The card was successfully overclocked to 1815 MHz boost, which is around 150 MHz faster than its standard speed. Additionally, memory could be overclocked to 1263 MHz, or 20.2 Gbps. Such outcomes were attainable despite the model’s locked TDP. This offered an additional 4 to 6 percent improvement in performance over the default settings.
Via ITHardware