We have a very very delicious rumor making the rounds today and it comes from a highly credible source. @kopite7kimi, the Twitter leaker responsible for pretty much all of the Ampere leaks, revealed a tidbit about the upcoming NVIDIA Ada Lovelace architecture (likely called NVIDIA ADA). Our colleagues over at 3DCenter extrapolated a lot of information from the die size, which Kopite appears to have more or less confirmed. While the source is highly reliable, we are still marking this post as a rumor because of the magnitude of the leak.

NVIDIA ADA GPU Leaked: Monster 64 TFLOPs GPU with 18432 CUDA Cores and 5nm process architecture

The Ada Lovelace architecture - which will likely only be referred to as NVIDIA ADA by the way - was recently leaked by Kopite (and confirmed by Videocardz) and we already seem to have the preliminary specifications of NVIDIA's upcoming GPU. As we mentioned in the original Ada article, Hopper appears to have been delayed for now (and along with it, NVIDIA's MCM ambitions). Thankfully, it seems that NVIDIA has kept its pedal to the metal and its Ada architecture, championed by the AD102 GPU will be an absolute beast. Given below is the original die size leak:

GA102 has a "7*6" structure.

Maybe AD102 will get a "12*6" structure.

— kopite7kimi (@kopite7kimi) December 28, 2020

The folks over at 3DCenter quickly extrapolated a ton of details (we have revised their TFLOP numbers to be a bit more conservative with a clock of 1.75 GHz) which Kopite confirmed:

And a larger cache. It looks like this.

— kopite7kimi (@kopite7kimi) December 28, 2020

For those that want all the information in one place, here is a table summarizing everything:

NVIDIA CUDA GPU (RUMORED) Preliminary: