When the original “SUPER” cards were introduced earlier this year NVIDIA was offering new configurations of existing silicon, with additional SMs – and thus more CUDA Cores – enabled, among other differences. This GeForce GTX 1660 SUPER launch is different, as it retains the identical configuration of a TU116 GPU found in the original (i.e. non-SUPER) version. This new GPU even runs at the same Base/Boost clocks as its predecessor!

So what makes this GTX 1660 SUPER, well, super? NVIDIA has upped the memory from 8 Gbps to 14 Gbps, which results in a jump from 192.1 GB/s to 336 GB/s available memory bandwidth on the same 192-bit bus. The big question then becomes, how memory bandwidth-constrained was the GTX 1660 to begin with? And can this new card come close to a GTX 1660 Ti? The gaming benchmarks to follow will answer that, and since we aren’t looking at any architectural or other significant changes let’s skip ahead to the important stuff: performance. (But first a quick look at specs.)