Navi 23 Might Be Based On AMD’s 2nd Generation RDNA Architecture, Internally Known As “NVIDIA Killer”

AMD recently announced its new RDNA architecture based Navi GPUs, and it appears the NVIDIA competitor is already deep in development of the second-generation RDNA. The new architecture reportedly supports high-bandwidth HBM2E VRAM. With their next-gen technology, AMD is attempting to achieve hardware-level integration on their GPUs to support real-time Ray Tracing in games. The rumored new high-end Navi 21 and Navi 23 cards are already being called the ‘NVIDIA Killer ’, and could arrive as early as next year.

Just a month after AMD launched their new RDNA architecture based Navi GPUs; there are indicators about the development of second-generation RDNA. The long-term prospect is quite exciting for consumers who are looking for powerful GPUs with serious clock-speeds and high bandwidth next-generation RAM. Needless to add, the new architecture will power the AMD Radeon RX Series graphics cards which are generally not meant for enthusiasts, but serious professionals and gamers.

AMD Navi 21 and Navi 23 ‘NVIDIA Killer’ GPUs With 2nd Generation RDNA Architecture And Ray Tracing:

AMD’s CEO recently confirmed during an investors call that they are indeed working on high-end, Navi based Radeon RX graphics cards. He did not reveal many details about their upcoming, and still in development, technologies. Hence the reports are still based on rumors, but they do fit well with AMD’s roadmap of long-term development. Incidentally, the new ‘NVIDIA Killer’ AMD GPUs are expected to launch in the second half of the next year primary because AMD would be introducing its second-generation RDNA architecture, built upon the 7nm+ process node, at that time.

The AMD Navi 21 and Navi 23 GPUs would most likely be the first commercial units that would be based on the new RDNA architecture. The recently launched high-end AMD Radeon RX 5700 series cards are good examples that showcase the technology’s power. Between the two new GPUs, it is the Navi 23 which could draw the most attention primarily because experts claim it will be this GPU that will be fused onto AMD’s next-gen enthusiast and high-end Radeon RX series graphics cards.

AMD Radeon R9 Fury X, Radeon RX Vega, and Radeon VII are powerful graphics cards, but they still fall behind NVIDIA’s high-end products. Interestingly, even similarly priced NVIDIA GPUs appear better in comparison. However, AMD is now seriously trying to accelerate its development and appears confident they could challenge NVIDIA’s flagship graphics cards. Needless to add, the NVIDIA GeForce RTX 2080 Ti is currently among the top-rated graphics cards, but AMD will have to work hard because NVIDIA could offer something better next year. Rumors indicate NVIDIA could launch the new GeForce 30 Series.

One of the most anticipated features within the upcoming high-end AMD graphics cards that are built with the 2nd generation RDNA architecture is the integral support for Ray Tracing. It is confirmed that AMD is working to ensure the new GPUs have hardware-level integration to support real-time Ray Tracing in games. Needless to add, once AMD manages to do so, its graphics cards will be at par NVIDIA’s RTX technology. Incidentally, NVIDIA achieved ray-tracing support through hardware-level integration last year itself. In fact, the currently available and popular GeForce RTX 20 series cards boast of Ray Tracing.

AMD Radeon On ‘7nm Navi RDNA’ GPU Specifications And Features:

The high-end AMD Navi GPUs should retain support for High-Bandwidth memory design just like the current flagship AMD products. Incidentally, the company currently solders GDDR6 memory on their mainstream RDNA based cards. However, for the next-generation RDNA architecture, AMD just might go for the new HBM2E VRAM.

The High Bandwidth Memory 2nd Evolution or HBM2E VRAM sports an 8-Hi stack configuration and utilizes 16 GB memory dies. The memory can be clocked at up to 3.2 Gbps. Stacked together, that’s a collective total bandwidth of 410 GB/s on a single and 920 GB/s with two HBM2E stacks. The DRAM has a 1024-bit wide bus interface which is identical to the current-generation HBM2 DRAM. These new ultra-high bandwidth memory modules are manufactured by Samsung.

The Korean tech giant claims their HBM2E solution, when stacked in 4-way configuration, can offer up to 64 GB memory at 1.64 TB/s of bandwidth. Needless to add, there are very few use-case scenarios that require or can even utilize such specifications. Barring some custom servers/HPC workloads, these high-end graphics products are more suited for serious gamers and professionals. Commercially available graphic cards could feature up to 32 GB memory with just two stacks. Incidentally, this configuration in itself packs twice as much memory as the AMD Radeon VII graphics card.

AMD has traditionally been the favorite choice of enthusiasts. The Radeon RX 5700 Series are quite popular and they confidently stand their ground against NVIDIA’s RTX SUPER Series. However, with the upcoming next-generation RDNA architecture and integrated Ray Tracing, the company is clearly attempting to appeal to consumers looking for serious performance.


Alap Naik Desai

A B.Tech Plastics (UDCT) and a Windows enthusiast. Optimizing the OS, exploring software, searching and deploying solutions to strange and weird issues is Alap's main interest.
Back to top button