The RTX 40 Series of GPUs, which include the RTX 4080 and the RTX 4090, were just unveiled by NVIDIA. They are powered by the newest Ada Lovelace Architecture, which is based on a brand-new 4nm process from TSMC and is designed specifically for GPUs. The RTX 40 Series GPUs also support new features like DLSS 3 and AV1 encoder.
NVIDIA RTX 4090
The NVIDIA RTX 4090 features an AD102-300 GPU with 16384 CUDA cores and a base clock of 2.23 GHz along with a boost clock of 2.52 GHz, which means a single-precision performance gain of 82.6 TFLOPS or 2.3x more than its predecessor, the RTX 3090.
Furthermore, the RTX 4090 will be equipped with 24GB GDDR6X memory with a peak bandwidth at 1 TB/s and the graphic card will require more power than 3090Ti by at least 100W with a default TDP of 450W, while needing a single 16-pin PCIe Gen 5 or 3×8-pin PCIe cables.
The RTX 4090 will be available starting October 12 at $1,599.
NVIDIA RTX 4080
The company also revealed two RTX 4080 models with differing GPU and memory specifications. The RTX 4080 16GB will have an AD103 GPU with 9728 CUDA cores, 16GB of GDDR6X RAM overall, clocked at 22.5 Gbps, and a TDP of 320W, according to the specifications.
The RTX 4080 12GB, on the other hand, features 7680 CUDA cores and an AD104 GPU with a boost frequency of 2.61GHz and a TDP of 285W.
In November, the RTX 4080 16GB and 12GB will be on sale for $1,199 and $899, respectively.