NVIDIA’s upcoming RTX 2060 revision to be a powerhouse with 1920 CUDA cores

The soon-to-be-released ‘upgraded’ RTX 2060 model is supposed to have more than 1920 CUDA cores. The card, currently known as the RTX 2060 12GB, is believed to have 2176 CUDA cores, the same as the RTX 2060 SUPER. Interestingly, the new SKU will feature a higher default TDP of 184W (compared to SUPER’s 175W), making this new news odder. On paper, the card is supposed to be almost identical to the 2060 SUPER, except for memory bandwidth: the 2060 SUPER model has a 256-bit bus, whilst the 12GB SKU has a 192-bit bus.

We were taken aback by this news, but we have subsequently confirmed it with three different sources. They’ve all been given the same instructions: The GPU has 2176 CUDA cores, 12GB GDDR6 14Gbps memory with a 192-bit bus, and a 184W TDP. According to two sources, the boost clock is set at 1650 MHz by default, with factory-overclocked models reaching 1690 Mhz.

NVIDIA RTX 2060 12GB will now be available on December 7th as a less expensive alternative to the RTX 3060 12GB (hopefully with better availability). It’s worth mentioning that at launch, the RTX 3060 12GB was around 8% quicker than the RTX 2060 SUPER. Given its almost equal specifications, the 2060 12GB may function similarly. The Turing model, on the other hand, will lack PCIe Gen4 and ResizableBAR support. In addition, the card’s raytracing performance will be reduced.

source