NVIDIA has a habit of creating many sub-variants of a particular GPU that often comes with no name change – bad for consumers, since there’s no clear way of telling them apart. Recently, Videocardz reported a new RTX 4070 variant has surfaced in the form of the GALAX RTX 4070 OC 2X, and the main difference simply comes down to slower VRAM.
RTX 4070 GDDR6
To recap, the NVIDIA GeForce RTX 4070 comes with AD104-250 silicon that features 5,888 CUDA cores and 12GB of GDDR6X memory, which is the faster version of GDDR6. This new variant will instead opt for a slightly different silicon codenamed AD104-251, though nothing is changed in terms of core counts. Instead, this new variant is paired with slightly slower GDDR6 memory rated at 20Gbps (as opposed to 21Gbps GDDR6X in the original model).
This further expands the RTX 4070 nameplate into a confusingly complicated lineup that features five distinct models with different performance characteristics (Ti SUPER, Ti, SUPER, standard, and GDDR6), plus another two that uses larger, binned-down silicon that otherwise behaves the same as their standard counterparts.
Should this new variant become widespread in the market, users should pay attention to the product name to differentiate between GDDR6 and GDDR6X modes, as a slight difference in memory bandwidth can result in varying performance in games. That said, it is unclear when this new variant will enter the market, or whether it comes with lower pricing.
Pokdepinion: At least this part of the specs is going to be labeled clearly, but I wouldn’t be surprised if people forgot to spot the ‘X’ letter.
