So far, Nvidia has revealed only some basic specifications for the GTX 1060 graphics card. Set to launch on July 19th, the only official confirmation for specifications are 1280 CUDA cores and a clock speed of about 1.7 Ghz. Today, we’re getting new information that details the rest of the core specifications.
From the block diagram, we’re now able to confirm the number of TMUs and ROPs are confirmed for 80 and 48 respectively. This a decent upgrade over the GTX 960 which only featured 64 TMUs and 32 ROPs. The core layout makes a lot of sense once you consider that Pascal is just an optimized Maxwell. The GTX 960 OEM variant had the exact same core layout of 1280 shaders, 80 TMUs and 48 ROPs. The GTX 1060 is able to challenge the GTX 980 though due to the much higher clock speeds made available by the die shrink to 16nm FinFETs and architecture tweaks.
With clocks speed ranging from 1506 to 1709 Mhz, the GTX 1060 will offer up to 4.4 TFLOPs of performance. This shouldn’t be able to pass the GTX 980, but likely just ahead of the RX 480. With just about a week and a half to go, it will be interesting to see how the match up between the two mainstream cards plays out.
