NVIDIA GeForce RTX 5090 and 5080: GPUs Ready to Boost Performance | Computer hardware parts list | Gpu hardware list | Hardware catalog | Turtles AI

The NVIDIA GeForce RTX 5090 and RTX 5080 GPUs promise a significant leap in performance. According to reliable sources, the RTX 5090 will be the flagship model with 21760 cores and 32 GB of VRAM, while the RTX 5080 will offer reduced specifications with 10752 cores and 16 GB of VRAM, creating a significant gap between the two cards.

Key points:

•  The RTX 5090 will have 21760 cores and 32 GB of GDDR7 VRAM.
•  The RTX 5080 will have 10752 cores and 16 GB of GDDR7 VRAM
•  The difference in cores between the two cards will be 51 percent
•  The expected TBP is 600 W for the RTX 5090 and 400 W for the RTX 5080

NVIDIA is now close to launching its generation of GeForce RTX 50 GPUs, known by the codename “Blackwell.” These graphics cards promise to bring significant performance improvements over the previous series. Among the most anticipated models are the GeForce RTX 5090 and RTX 5080, whose technical specifications were recently unveiled. The GeForce RTX 5090 promises to be the top-of-the-line GPU, built on the PG144/145-SKU30 PCB design and equipped with a GB202-300-A1 core. This card will have 170 SMs enabled out of a total of 192, for a total of 21760 active cores, representing an 11.4 percent reduction from the chip’s maximum potential. This is slightly higher than the 11.1 percent reduction found in the RTX 4090 compared to the full AD102 die. On the memory front, the RTX 5090 will feature as much as 32 GB of GDDR7 VRAM, arranged on a 512-bit interface. The first GDDR7 memories will offer data transfer rates between 28 and 32 Gbps, translating into a total bandwidth of 1.792 or 2.00 TB/s. Power consumption, or TBP (Total Board Power), will be 600 W, although this value does not necessarily represent actual power consumption during daily use or gaming, which is usually lower. Despite this high power, the RTX 5090 will feature a 2-slot heatsink in the Founders Edition variants, demonstrating a balance between power and design.

On the other hand, the RTX 5080 will be a more downsized version of its big sister. Based on the PG144/147-SKU45 PCB, it will incorporate the GB203-400-A1 GPU die, making full use of its 84 SMs and 10752 cores. However, this configuration represents a 51 percent reduction from the RTX 5090, a much more pronounced difference than the 40 percent gap that existed between the RTX 4090 and RTX 4080. In addition, the RTX 5080 will be equipped with 16 GB of GDDR7 VRAM on a 256-bit bus interface, with memory bandwidth varying between 896 GB/s and 1024 GB/s. The TBP of the RTX 5080 will also be lower, set at 400 W, a 25 percent increase over the previous generation, but as always, the actual figures may differ. 

The significant difference in specifications between the high-end GPUs, particularly between the RTX 5090 and RTX 5080, could have implications for NVIDIA, as it did with the RTX 4080, which received criticism for its price-performance ratio. The price of the RTX 5080 could play a crucial role in determining the success of this card, with many users hoping for a figure below $1000, although the odds are low given NVIDIA’s monopoly in the high-end market segment. However, it must be considered that the leaked specifications could still undergo changes before the official launch, expected by the end of the year or early next year. In any case, more details will emerge in the coming months.

We remain waiting for official confirmations to understand the actual impact of these new GPUs on the market.