NVIDIA Quietly Refreshes GeForce RTX 2070, RTX 2060 Mobile GPUs With Increased Performance
Apparently, those aren't the only changes that NVIDIA is making to its mobile GPU lineup, as a representative for German OEM Schenker spilled the beans on "refreshed" versions of the GeForce RTX 2060 and RTX 2070 mobile GPUs. Rather than relying on additional CUDA cores to increase raw processing power, this hardware refresh takes a more measured approach.
In this case, NVIDIA is pairing the GPUs with low-voltage GDDR6 memory, which operates at 1.25V instead of 1.35V. The new memory should result in lower power consumption and less heat that is generated by the GPU module inside the laptop. This change has the added benefit of giving NVIDIA some additional headroom to adjust the boost clocks on the GPU while adhering to the preset total graphics power (TGP) rating.
The GeForce RTX 2060 had previously been rated for 80W TGP, but that has now been bumped to 115W. The GeForce RTX 2070, however, retains its 115W TGP. Combined with the more efficient GDDR6 memory, the GeForce RTX 2070 Refresh (N18E-G1R) is said to have seen its base and boost GPU clocks increased to 1305MHz and 1485MHz respectively according to Notebook Check. There is no word at this point on the revised clock speeds for the refreshed GeForce RTX 2060 (N18E-G1-B)
According to Schenker rep "Tom", NVIDIA’s GeForce RTX "refresh cards exceed their non-refresh predecessors by significant margins with the same power consumption." The assumption is that NVIDIA has been able to improve yields on the chips as well, but there's no confirmation on that rumor.
If the Schenker rep's commentary is accurate, we'll begin to see the first benchmarks for the refreshed GeForce RTX 2060 and RTX 2070 mobile GPUs on April 15th.