RTX 4090

Comparing the RTX 3090 Ti and 3080 Ti to the RTX 4090 and 4080

a generational gap

During its GTC 2022 keynote, Nvidia finally unveiled the RTX 40-series of graphics cards after months of rumours and teasers. The RTX 40-series, which uses the new Ada Lovelace graphical architecture, will replace the RTX 30-series, which made its debut in 2020.

In less than a month, the RTX 40-series will go on sale, so let’s look at the specs, explore the unique features, and compare the performance of the initial RTX 40-series GPUs to the RTX 3090 Ti and RTX 3080 Ti.

Brief comparison of the RTX 40 and the RTX 3090 Ti and 3080 Ti
The RTX 40-series will feature a number of changes from the RTX 30-series, as shown in the infographic above. The RTX 40-series uses GDDR6X memory throughout, but the specifications are not all that different from those of the Ti versions of the RTX 3090 and 3080.

According to Nvidia, the RTX 4080 variations would perform up to twice as well as the RTX 3080 Ti. In contrast, it is asserted that the RTX 4090 Ti is two to four times quicker than the RTX 3090 Ti.

The GPUs’ support for DLSS, Nvidia’s supersampling technology, as well as hardware-based ray tracing are other parallels that are worth mentioning. DLSS is supported by the RTX 30-series, however there are significant differences that we’ll go through in the following section.

Two RTX 4080 models are available, one with 16GB of GDDR6X memory and a suggested retail price of $1,199, and the other with 12GB of GDDR6X memory and a suggested retail price of $899. This would not be the first time Nvidia had done this; the company had previously launched two separate models of its RTX 3080 graphics cards, the first of which had 10GB of RAM and the second of which had 12GB.

There are more differences between the two RTX 4080 models besides memory (and price). The 16GB model uses at least 750W of power and contains additional RT, Tensor, and CUDA cores. The 12GB variant, in comparison, has a faster base and boost clock and needs at least 700W of power.

Architecture Comparison: RTX 40 vs. RTX 3090 Ti and 3080 Ti
The RTX 30-series is based on Samsung’s 8nm technology and Nvidia’s Ampere graphics architecture. The RTX 40-series, on the other hand, is constructed using TSMC’s 4nm N4 technology and the new Ada Lovelace architecture.

The third-generation Tensor and second-generation RT cores are used in the Ampere architecture, which is the second generation of RTX. The third generation of RTX, known as the Ada Lovelace architecture after the mathematician and author, uses fourth-generation Tensor cores and third-generation RT cores. Additionally, each Lovelace RT and Tensor core is more potent than each Ampere RT and Tensor core, as you could anticipate from a new generation of microarchitecture.

GeForce Beyond vs. Microsoft Flight Simulator DLSS 3 Comparison Trailer 0:55 Autoplay mode: On GeForce Beyond Comparison Trailer for Microsoft Flight Simulator DLSS 3

The third generation of DLSS is only supported by the Lovelace architecture, which is also the sole graphical architecture. DLSS 3 is “driven by the new fourth-generation Tensor Cores and Optical Flow Accelerator,” according to a blog post by Nvidia, which is only available in the RTX 40-series.

These comparisons, of course, only tell the tale on paper; until we have the chance to benchmark the RTX 40-series ourselves, we won’t properly understand how the GPUs compare. The new manufacturing method and visual architecture, however, give hope that there will be a noticeable performance increase between the two generations of GeForce graphics cards.

Total
0
Shares