Nvidia just unleashed the full potential of Ada Lovelace, and the RTX 4090 Ti GPU has 70% more CUDA cores than the RTX 3090 Ti. The same boost applies to the AD102 chip’s Tensor and ray tracing cores, but its clock speeds are in line with the recently announced GeForce RTX 4090.
During Nvidia’s virtual press event, the company revealed the maximum specifications for the RTX 4000 GPU, highlighting the differences between Lovelace and Ampere. Although the RTX 4090 lacks the full AD102 chip, upgraded variants like the RTX 4090 Ti can include all 18,434 CUDA cores, 144 Gen 3 RT cores, and 576 Tensor Gen 4 cores available.
| Nvidia AD102 graphics card | |
| Video memory (GDDR6X) | 24GB |
|---|---|
| speed up the clock | 2.52GHz |
| Basic Clock | 2.23GHz |
| CUDA kernels | 18,434 |
| Ray tracing kernels | 144 (third generation) |
| tensor nuclei | 576 (fourth generation) |
| tire width | 384 bit |
By comparison, the full Ada Lovelace GPU has 71% more cores than its predecessor Ampere GA102. However, choosing the RTX 4090 Ti doesn’t seem to give you faster speeds than the RTX 4090 as they both have a maximum clock speed of 2.5GHz.
On Tuesday, Nvidia officially announced the RTX 4080 and RTX 4090, two cutting-edge graphics cards powered by Ada Lovelace’s new architecture. Also, the green team introduced a new version of its AI scaler, but it will only be able to use DLSS 3.0 on RTX 4000 GPUs.
Nvidia hasn’t announced a lead candidate for an all-GPU AD102 graphics card, but it will potentially come later in the RTX 4000-series lifecycle. There’s also no sign of an RTX 4060 or RTX 4070 at the moment, but two RTX Se expect 4080 models. The 12GB VRAM variant is already causing a stir among enthusiasts as it lags behind its 16GB counterpart in terms of specs and performance.
Source : PC Gamesn
