Tipster @CyberCatPunk released what appears to be a spect sheet for NVIDIA’s upcoming GeForce RTX 3080 Ti. The leaked spec sheet claims that the upcoming Ampere flagship will deliver 21 TFLOPs of raw compute, 50 percent higher than NVIDIA’s current performance king the GeForce RTX 2080 Ti.
This also has major implications with regards to AMD’s upcoming Big Navi GPU. Purported leaks indicated that, as of now, Big Navi only delivers 10-15 percent higher performance than the GeForce RTX 2080 Ti. If the 3080 Ti spec list holds true, this might put AMD’s high-end performance at a significant disadvantage.
The spec sheet claims that the GeForce RTX 3080 Ti will feature 5376 CUDA cores running at up to 2.2 GHz. Current Turing cards top out at roughly 2.1 GHz so this is a notable bump to clock speeds.
The spec sheet also claims that the RTX 3080 Ti will leverage 18 Gbps Samsung GDDR6 to deliver 863 GB/s of memory bandwidth. Together with its prodigious compute capabilities, this should be enough to deliver an uncompromised 4K experience, and even playable 8K frame rates.
The leak did also mention that, despite being built on the efficient 7nm process, the RTX 3080 Ti will have a 320W TDP. If true, this would mean that Comet Lake S i9-10900K builds would require an eye-watering level of power, otherwise only seen in top-end SLI setups.
The leak isn’t confirmed so we’d take some of these specs with a grain of salt. We will update you, however, as soon as we get validated, official details on Ampere.