Given the lackluster reception of real-time ray tracing on GeForce Turing-powered graphics cards, Nvidia may be modifying its marketing plan to release GTX 11-series graphics cards without the 20-series' RT cores. While this was a mere rumor a week ago, Lenovo's listing today seems to confirm the GTX 11-series' existence.
Like its RTX 20-series siblings, the future GTX 11-series will employ Nvidia's latest Turing graphics card architecture. However, they will not share the same silicon. The GeForce RTX 2060 will purportedly utilize the TU106 die, whilethe GeForce GTX 1160 will employ the TU116 die instead. Both chips are manufactured by TSMC under its 12nm FinFET node, of course.
|GeForce RTX 2060||GeForce GTX 1160|
|Architecture (GPU)||Turing (TU106)*||Turing (TU116)|
|Base Clock Rate||1320 MHz*||?|
|GPU Boost Rate||1620 MHz*||?|
|Memory Capacity||6GB GDDR6*||6GB GDDR6*|
|Memory Clock||14 Gbps*||14 Gbps*|
|Memory Bandwidth||336 GB/s*||336 GB/s*|
|Transistor Count||10.8 billion*||?|
|Die Size||445 mm²*||?|
At this moment in time, there is no concrete evidence on the level of differences between the two graphics cards. Early speculation is that the GeForce GTX 11-series will lack the Tensor and RT cores. Both graphics cards should feature up to 1920 CUDA cores and 6GB of GDDR6 memory clocked at 14 Gbps. The RTX variant could sport 240 Tensor cores for AI and 30 RT cores for ray tracing. It also reportedly comes with a 1,320MHz base clock and 1,620MHz boost clock.
Thanks to Lenovo's recent leak, we can be fairly certain that the GeForce GTX 1160 will be offered in 3GB and 6GB memory configurations. This shouldn't really come as a surprise as Nvidia released the previous GeForce GTX 1060 with 3GB and 6GB as well, and then later launched a region-specific variant with 5GB.
The GeForce RTX 2060 is expected to be announced on January 15, 2019.