New and long-awaited, next-generation NVIDIA GeForce GTX 1180 graphics card has been added to the venerable TechPowerUp database confirming some specifications. The data is about an engineering sample, and we can see things like improved clock speed so that we would be entering the final stages of production. Rumours are that it could be presented at the Computex 2018.
NVIDIA GTX 1180 will be manufactured in a 12 nm FinFET process
The data entered is mostly identical to the data that had been filtered. The NVIDIA GTX 1180 graphics card will be built with TSMC’s 12nm FinFET process, which will deliver significant power efficiency improvements. It will have precisely 3584 CUDA cores divided into 28 SM, 64 ROPs and 224 TMUs. According to the same entry, the memory in question is the GDDR6 variant with up to 16 GB of DRAM memory.
The clock speed for the memory is 12 GHz effective, which is one step ahead of Pascal. The GPU core clock is listed at 1405 MHz and can reach 1582 MHz in Turbo. The pixel rate is 101.2 GP pixels/s, and the texture rate is 354.4 GTexels/s. The maximum floating point performance will be around 13 TeraFlops.
The TDP will be 200 W which can be powered by a 1×6 pin and 1×8 pin configuration.
GeForce GTX 970s are in short supply
Turing appears to be mainly a process of reducing and optimizing the Volta architecture and will bring significant performance and energy efficiency to the next generation graphics card.