After weeks of teasing, the “Ada Lovelace” generation of Nvidia’s latest computer graphics cards, the RTX 4000 GPU, are here. Nvidia CEO Jensen Huang introduced two new models on Tuesday: rtx 4090, starting at $1,599, and RTX 4080Which will launch in two configurations.
The pricier card, launching on October 12, is in the same high-end range as Nvidia’s 2020 megaton RTX 3090 (previously designated by the company as its “Titan” product). The 4090’s increase in physical size will demand three slots on your PC build of choice. The specs are indicative of the highest-end GPU: 16,384 CUDA cores (up from the 3090’s 10,496 CUDA cores) and a boost clock of 2.52 GHz (up from 1.695 GHz on the 3090). Despite the improvements, the card still performs within the same 450 W power envelope as the 3090. Its RAM allocation will be on 24GB GDDR6X memory.
This jump in performance is driven by Nvidia’s long-standing leap to TSMC’s “4N” process, a new generation of 5nm chips that deliver a massive efficiency jump from the previous Ampere generation’s 8nm process .
Meanwhile, the RTX 4080 will follow in November in two SKUs: a 12GB GDDR6X model (192-bit bus) starting at $899, as well as a 16GB GDDR6X model (256-bit bus) starting at $1,199. Depending on how different the specifications are between these two models, Nvidia is launching two completely different chipsets under the same “4080” banner; Traditionally, so much Nvidia hardware differentiation has come with different model names (ie the previous generation 3070 and 3080). We are awaiting confirmation from Nvidia whether the 4080 models share the chipset.
Both of these will also include more CUDA cores (9,728, up from the RTX 3080’s 8,704) in price, boost clock (2.51 GHz, up from the 3080’s 1.71 GHz), and power draw (320 W, similar to the 3080 but higher than the lower) . -Memory 4080’s 285 W). At least for this generation, the Nvidia 4080 is offering 12GB of memory as a baseline—effectively addressing a major criticism faced with the increasingly memory-hungry RTX 3000 generation of GPUs.
Both new models feature iterative updates to Nvidia’s proprietary RTX chipset elements—RT Core and Tensor Core. In addition, Nvidia has announced updated processes for both real-time ray tracing in 3D graphics and its Deep Learning Super-Sampling (DLSS) upscaling system. The former would be augmented on Lovelace GPUs with two new types of hardware units: an “Opacity Micromap Engine”, meant to double the raw ray-tracing performance, and a “Micromesh Engine” to increase the amount of geometric coverage. On the rendering front “without the storage cost”.
The latter is now joined by a new version: DLSS 3, which appears to be an RTX 4000-series exclusive feature. According to Huang, the system promises to “generate new frames”. [of gameplay] Effectively enhances the performance of both CPU and GPU, without involving the game.” Frame-by-frame reconstruction of graphics in real time is designed to address the issues found in image reconstruction techniques that may not necessarily model the motion vectors of in-game elements such as particles. This new method, along with techniques established by prior DLSS generations, was used to reconstruct 7/8 of a scene’s pixels – thus dropping the computational demands on both the GPU and the CPU. If it works as promised, DLSS can deliver a significant improvement over the pixel-by-pixel process that DLSS and its competitors – AMD K has been so successful in both FSR 2.0 and Intel’s upcoming XeSS.
On top of those proprietary systems, Nvidia’s latest GPUs will apparently rely on a new process that Huang calls “shader execution reordering.” While this would improve the performance of raw rasterization, Huang’s brief description of the system relies largely on the computationally expensive workload of ray tracing. The system will run “2-3 times” the ray-tracing performance of the company’s previous Ampere generation GPUs.
As part of today’s announcements, Nvidia presented a few titles with upcoming Nvidia RTX-specific customizations, and the biggest is arguably Microsoft Flight Simulator 2020, When combined with DLSS 3’s image reconstruction system, MSFS Some of the game’s busiest scenes were displayed, with frame rates easily reaching the 100s (though an unnamed Lovelace GPU and raw rasterization of similar scenes on PC also showed solid performance, given that that game and how power-hungry its expansive cityscapes can be) The gallery above is captured in 4K, so you can see how DLSS 3 works fine compared to the obvious combination of raw pixels and standard Temporal Anti-Aliasing (TAA) in the older system. You may want to click and zoom in to see how the pixels handle the details.
Huang also tried to sided with the PC modding scene, by demonstrating how a new Nvidia-developed toolset, called RTX Remix, could be applied to a number of classic games whose modding capabilities are extensive. The best results came from a before and after exposure that . applies to The Elder Scrolls III: Morrowind, which uses RTX GPU tensor cores to programmatically update the game’s texture and physical material properties. (see above, or Visit Nvidia’s site for even more before and after comparisons.) Similar results in a free . are expected from Portal 1 The DLC package that Nvidia will release in November for fans to apply to that PC gaming classic. We’ll be curious to see a face-to-face demonstration of Nvidia’s automatic modding results in a comparison of the best stuff from more than a decade of community development.