Different sites recently reported on the presence of two different SKUs of each Turing GPU. It was noticed a while ago that Nvidia is using two different GPU hardware IDs for the same Turing GPU. One of them is the standard SKU and the other one an OC version.

So you must be wondering what’s all this about and what difference does it make. To clarify this, let’s take GeForce RTX 2080 Ti as an example. The GPU used for this is a TU 102. So, Nvidia offers two different chips based on TU 102, one of them is TU 102-300 and the other one TU 102-300-A. The A model is the OC SKU.


So, a board partner is allowed to factory tweak the GPU only if it is using the A model i.e the OC SKU. Since the non-A variant would then operate at reference clock speeds, it would be a cheaper blower cooler style product. A more high-end GPU would probably be the A variant.

You can still overclock the non-A models yourself using tools like Afterburner or any other tool of your choice. But it is highly probable that the overclocking capabilities of the non-A model would be less than that of the A model.