NVIDIA officially introduced the GeForce RTX 40 GPUs featuring a special keynote from its CEO, Jensen Huang. The new series consists of three SKUs based on the new “Ada” architecture.
NVIDIA’s GeForce RTX 40 lineup features the flagship RTX 4090 and two RTX 4080 models with varying memory specs.
The new architecture introduced 3rd Gen RT Core and 4th Gen Tensor Core. It also supports DLSS 3 upscaling technology, and the dual AV1 media encoders offer up to 8K 60 FPS recordings. The RTX 40 GPUs will use the PCIe Gen4 x16 interface.
NVIDIA RTX 4090
The flagship RTX 4090 features AD102-300 GPU with 16384 CUDA cores and a boost clock measuring up to 2520 MHz. The SKU comes with 24GB of GDDR6X memory and a bandwidth of 1 TB/s.
The RTX 4090 features 76 billion transistors and the latest TSMC 4N process technology. Although its default TDP is 450W, the card will demand at least 100W or even more power than the RTX 3090 Ti.
The processor will offer 82.6 TFLOPS in single-precision performance, which is 2.3 times more than its predecessor, the RTX 3090.
The RTX 4090 Founders Edition flaunts a triple-slot design with a 16-pin power connector.
NVIDIA has announced multiple designs from its board partners: ZOTAC, Inno3D, Colorful, Palit, Gainward, Gigabyte, MSI, and ASUS. However, EVGA did not make the cut.
The RTX 4090 GPU will be launched on October 12, priced at $1599.
RTX 4080 (16GB & 12GB)
NVIDIA introduced two GPUs with the same model number but different specs– the RTX 4080 12GB and RTX 4080 16GB.
Contrary to what was reported earlier, the RTX 4070 is now called the RTX 4080 12 GB. This model features the AD104-400 GPU with 7680 CUDA cores. It supports 12GB of GDDR6X memory and has a default TDP of 285W.
The RTX 4080 16GB is paired with the AD103-300 GPU and 9728 CUDA cores. It supports 16GB of GDDR6X memory, with a speed of 22.5 GB/s. The default TDP is set at 320W, identical to the RTX 3080.
The MSRP for the RTX 4080 12GB and the 16GB models are $899 and $1199, respectively. The cards are scheduled to launch in November.
Here’s the official specs list:
| RTX 4090 | RTX 4080 12GB | RTX 4080 16GB | |
| Board-SKU | PG139-SKU330 | PG139-SKU331 | PG141-SKU360 |
| GPU | AD102-300 | AD104-400 | AD103-300 |
| SMs | 128 | 60 | 76 |
| CUDA Cores | 16384 | 7680 | 9728 |
| Boost Clock | 2520 MHz | 2610 MHz | 2505 MHz |
| Memory | 24GB GDDR6X | 12 GB GDDR6X | 16GB GDDR6X |
| Memory Bus | 384-bit | 192-bit | 256-bit |
| Memory Speed | 21 Gbps | 21 Gbps | 23 Gbps |
| Bandwidth | 1008 GB/s | 504 GB/s | 736 GB/s |
| Default TDP | 450W | 285W | 320W |
| Max TDP | 660W | 366W | 516W |
| MSRP | $1599 | $899 | $1199 |
About Nvidia
NVIDIA Corporation is an American multinational technology company incorporated in Delaware, based in Santa Clara, California. They design graphics processing units (GPUs) for the gaming and professional markets, as well as system on chip units (SoCs) for the mobile computing and automotive market.
Best known for the “GeForce” lines of GPUs, they are a direct competitor to AMD’s “Radeon” series. NVIDIA has also expanded its offerings with its handheld game consoles Shield Portable, Shield Tablet, and Shield Android TV and its cloud gaming service GeForce Now.
No Comments on NVIDIA’s RTX 40 Series Features Two Memory Variants for the RTX 4080