Please note that this post is tagged as a rumor.
Alleged full AD102 GPU with 160+ FPS in Control at 4K
Just a day after first performance leaks on GeForce RTX 4090, we now have rumors on the full AD102 GPU.
The full AD102 GPU should not be confused with RTX 4090 non-Ti. The full GPU features as many as 18432 CUDA cores which is 12.5% more than alleged RTX 4090 specs. This might be the flagship RTX 4090 Ti model which was part of a very early cooler design leak a few months ago. However, chances are that NVIDIA will be releasing a new TITAN this time around, so it is unclear what model this is specifically.
According to @XpeaGPU who claims to be ‘a hunter of fake leaks’, full AD102 GPU offers more than 160 frames in game Control. This would be achieved at 4K resolution with raytracing and DLSS enabled. Furthermore, the sample had a very high power draw, which would align with previous rumors that full AD102 could consume as much as 800W.
3090 in Control with the same settings does 72 FPS on Avg so 2.2x gain pic.twitter.com/UdK2kCEXhO
— Hassan Mujtaba (@hms1193) July 19, 2022
That is indeed a claim with many variables missing. For starters, we do not know the exact game settings and whether Ada uses a new generation of Tensor cores (which would boost DLSS). Furthermore, the full specs of the graphics card sample are unknown (TDP or clocks).
Nevertheless, we know how fast the current flagship RTX 30 series performs in this game, and the results are impressive. That would be at least 2x better performance (depending on the DLSS profile).
It will probably take time to verify the claims from XpeaGPU, who is relatively new to the leak scene.
|RUMORED NVIDIA GeForce RTX 40 Series Specs|
|VideoCardz.com||Full AD102 (ADA TITAN?)||GeForce RTX 4090||GeForce RTX 4080||GeForce RTX 4070|
|Architecture||Ada (TSMC N4)||Ada (TSMC N4)||Ada (TSMC N4)||Ada (TSMC N4)|
|Memory||TBC||24 GB G6X||16 GB G6X||10 GB G6|
|Memory Speed||TBC||21 Gbps||21 Gbps||18 Gbps|
|Launch Date||TBC||October 2022||November 2022||December 2022|
Source: @XpeaGPU, @harukaze5719