Phil Hayton >Published Sep 21, 2022 at 10:58 ET >Gaming hardwareNvidiaNvidia just showed Ada Lovelace's full potential, and the RTX4090 Ti GPU could boast a 70% greater CUDA core count that the RTX3090 Ti. The same increase is seen in the AD102 chip’s ray tracing, Tensor cores, and its boost clock speeds match those of the GeForce RTX 4090.
Nvidia revealed the maximum RTX 4000 GPU specifications during a virtual press conference. This highlighted the differences between Lovelace & Ampere. Although the RTX 4090 doesn't have an AD102 chip, souped up variants such as the RTX 4090Ti might include all 18,434 CUDA cores and 144 third-gen RT Cores. There are also 576 fourth-generation Tensorcores.Nvidia AD102 GPUVRAM(GDDR6X),24GBBoost clock2.52GHzBaseclock2.23GHzCUDA Cores18,434Ray Tracing cores144 (Third generation)Tensor cores576 (foruth generation)Bus Width384–bit
The Ada Lovelace GPU has 71% more cores compared to its Ampere GA102 predecessor. Choosing an RTX4090 Ti will not give you faster speeds than the RTX4090. Both come with a 2.5GHz boost clock ceiling.
Nvidia announced Tuesday the RTX 4080, and RTX 4090 next-generation graphics cards. They feature the Ada Lovelace architecture. The green team also revealed a new version its AI upscaler. However, you can only use DLSS 3.0 with RTX 4000 GPUs.
Nvidia has not yet announced a best graphics cards contender equipped with a full AD102 GPU. However, it may arrive in the RTX4000 series' life cycle. The RTX 4060 and RTX 4070 are still not available, but Nvidia has promised two RTX 4080 models in November. Fans are already talking about the 12GB VRAM flavour, which is slower than its 16GB counterpart in terms both specs and performance.
Did you miss our previous article...