NVIDIA RTX 4090 Spotted in TSE, Scores 20,000 Points at 3.01GHz

Just a few days ago, a render came up showcasing the RTX 4090 from NVIDIA’s yet-to-launch Lovelace series. Today, the RTX 4090 has been allegedly tested in TSE (Time Spy Extreme) showing impressive results, in line with our original expectations.
Meet the RTX 4090
On launch, the RTX 4090 may be the flagship with a -Ti refresh possibly launching next year. Lets first get to the specs, the RTX 4090 features the AD102 GPU along with 16384 Cuda cores. Paired with a 382/384-bit memory bus the 24GB of 23Gbps memory amounts to a massive ~1100GB/s of effective bandwdith.
A Huge Performance Increase
Today over at Chiphell, a screenshot came up possibly showing the RTX 4090 from NVIDIA. The GPU scores a massive 20,192 points which is much higher than the initial ~19,000 points expectation. This is insanely high (Almost 3x the RTX 3080).
The total score is slightly lower as Videocardz mentions that an i5-12400F was used, which must’ve decreased the results by a bit.

The temperature at the time of silent frequency running is 30 degrees at room temperature.
This card is air-cooled, and the radiator is very large. It is estimated that the middle tower may not be able to fit, and it is impossible to take pictures. It looks like this is designed to look like 600-800w, but the default tdp is very low and only 450w, it is estimated that non-guild will have 800w version— 烟花易凉, Chiphell via Videocardz
The leaker also shared a screenshot showcasing the ‘temperature’ and ‘GPU clocks’. The GPU clocks in at a massive 3.01GHz, however, the temperatures are at a rather cool 55.07C. The leaker did mention that an extremely large radiator was used, possibly making up for such temperatures. The TDP is rated at just 450W.

Expected Specifications:
SKU | Chip | FP32/CUDA | Max Clock | Cache | Memory Bus | VRAM | Memory Spec | Speed (Gbps) | TDP | |||
RTX 4000 Titan | AD102-450 | 18176 | 3.0GHz+? | 96MB? | 384/382-bit | 48GB | GDDR6X | 24 | ~800W | |||
RTX 4090 Ti | AD102-350 | 18432 | 3.0 GHz? | 96MB | 382-bit | 24GB | GDDR6X | 24 | 600W? | |||
RTX 4090 | AD102-300 | 16384 | 3.0GHz+ | 96MB | 382-bit | 24GB | GDDR6X | 23 | 450W | |||
RTX 4080 Ti | AD102 | 14848? | 2.7 GHz? | 80MB? | 320-bit | 20GB | GDDR6X | 23 | 420W | |||
RTX 4080 | AD103-300 | 10240 | 3.2 GHz? | 64MB | 256-bit | 16GB | GDDR6X | 23 | 320W | |||
RTX 4080 (Variant 1) | AD103-300-A1 | 9728 | 3.1GHz? | 48MB | 256-bit | 16GB | GDDR6X | 23 | 320W | |||
RTX 4080 (Variant 2) | AD103-300 | 9728 | 3.1GHz? | 48MB? | 256-bit/192-bit | 12GB | GDDR6X | 23? | ||||
RTX 4080 (Variant 2 by Kopite) | AD104-400? | 7680 | 3.0GHz+? | ? | 256-bit/192-bit | 12GB | GDDR6X | 21 | 285W | |||
RTX 4070 Ti | AD104-400? | 7680 | 3.0GHz? | ? | 192-bit | 12GB | GDDR6X | 21 | 300W? | |||
RTX 4070 (Original) | AD104-275? | 7680 | 2.7GHz+ | ? | 192-bit | 12GB | GDDR6X | 21 | 285W | |||
RTX 4070 (New) | AD104-275? | 7168 | ? | ? | 160-bit | 10GB | GDDR6X | 21 | 250W |